1908 J Hygiene - The Prevention of Compressed-Air Illness (Boycott)

342
THE PEEVENTION OF COMPRESSED-AIR ILLNESS.

BY A. E. BOYCOTT, D.M.,
G. C. C. DAMANT,
Lieut, and Inspector of Diving, K.N.,
AND J. S. HALDANE, M.D., F.R.S.
{From the Lister Institute of Preventive Medicine.}
[With 7 Figures and 3 Plates.]
CONTENTS.
FAOE
Introduction . . . . . . . . . . . . 343
Part I. Theoretical.
A . T h e r a t e of s a t u r a t i o n o f t h e b o d y w i t h n i t r o g e n d u r i n g e x p o s u r e
to compressed a i r . . . . . . . . 345
B . T h e r a t e o f d e s a t u r a t i o n of t h e b o d y w i t h n i t r o g e n d u r i n g a n d
after decompression 350
C . T h e l i m i t s of s a f e t y i n d e c o m p r e s s i o n . . . . . . 355
D . P r a c t i c a l m e a s u r e s for a v o i d i n g c o m p r e s s e d - a i r i l l n e s s :
(1) D i v i n g 358
(2) W o r k i n c a i s s o n s a n d t u n n e l s 372
Part I I . Experimental.
1. A p p a r a t u s 377
2 . C h o i c e of e x p e r i m e n t a l a n i m a l s a n d r e s u l t s w i t h l a r g e a n d s m a l l
animals 378
3 . E e s p i r a t o r y e x c h a n g e of g o a t s 380
4 . M e t h o d of c o n d u c t i n g t h e e x p e r i m e n t s . . . . . 382
5. S y m p t o m s observed i n g o a t s 385
6. R e s u l t s of t h e g o a t e x p e r i m e n t s . . . . . . . 394
7. I n d i v i d u a l v a r i a t i o n a m o n g t h e e x p e r i m e n t a l a n i m a l s . . 403
8 . P a t h o l o g y of c a i s s o n d i s e a s e i n g o a t s . . . . . . 410
Summary 424
Appendix I. H u m a n experiments i n the pressure chamber . . . 426
„ I I . H u m a n experiments i n deep diving . . . . . 429
,, I I I . P r o t o c o l s of s o m e g o a t e x p e r i m e n t s . . . . . 436
ii I V . T a b l e s f o r d e c o m p r e s s i o n of d i v e r s . . . . . 442
A. E. BOYCOTT, G. C. C. DAMANT AND J. S. HALDANE 343
INTRODUCTION.
MEN who have been working in compressed air, as in diving, pre-

paring foundations of bridges, etc. under water, or making tunnels or
shafts through water-bearing strata, are liable on their return to
atmospheric pressure to a variety of symptoms generally known as
" diver's palsy " or " caisson disease," but which may more conveniently
be designated " compressed-air illness." It was shown experimentally
by Paul Bert1 that these symptoms are due to the fact that gas (chiefly
nitrogen) which goes into solution in the blood and tissues during
exposure to compressed air is liberated in the form of bubbles on too
rapid decompression, and produces local or general blockage of the
circulation or other injury. Subsequent investigations, for an account
of which we must refer more particularly to the treatise on the subject
by Heller, Mager and v. Schrotter2 and to recent papers by Hill and
McLeod3 and Hill and Greenwood4, have confirmed and extended Paul
Bert's conclusions.
It was pointed out by Paul Bert that by means of very slow
decompression the symptoms of caisson disease could be avoided, but
his experiments were not sufficient to furnish data as to what rate of
decompression would be safe. Nor has subsequent human experience
in engineering undertakings solved this problem; and the risks at-
tending work in compressed air at excess pressures of over 1^ to 2
atmospheres are notorious. Heller, Mager and v. Schrotter have
endeavoured to formulate rules as to safe decompression; and they
express the belief that perfectly uniform decompression at the rate
of 20 minutes an atmosphere would always be safe. Following this
rule, which is based on a calculation, Hill and Greenwood decompressed
themselves, without any serious symptoms, after short exposures at
excess pressures of as much as five and even six atmospheres.
Although the rules formulated by the above-mentioned observers
constituted a distinct step in advance, it appeared to us that, for
reasons which will be explained below, there were grave doubts as
1
La Pression Barometrique, 1878.
2
Luftdruckerkrankungen, 1900; also v. Schrotter, Der Sauerstoff in der Prophylaxie
und Therapie der Luftdruckerkrankungen, 2nd edition, 1906. The former work contains
a very full abstract of all previous investigations on the subject.
3
This Journal, vol. in. (1903), p. 401 (and references there given) : see also Becent
Advances in Physiology, 1906, pp. 233—255.
4
Proceedings of the Royal Society, vol. LXXVII. p. 442, 1906; vol. LXXIX. p. 21,1907;
also British Medical Journal, July 7th, 1906, Feb. 16th, 1907, June 22nd, 1907.
22—2
344 The Prevention of Compressed-air Illness
to the safety of their recommendations, and particularly as to whether
uniform decompression is desirable. The need for framing definite
rules as to safe decompression in the shortest possible time presented
itself in a very definite form in connection with the work of the
Admiralty Committee on Deep Diving1, of which one of us was a
member. Our investigation, which was planned with the more
particular object of furnishing information required for securing the
safety of divers ascending from deep water, was rendered possible by
the gift to the Lister Institute by Dr Ludwig Mond, F.R.S. of a large
experimental steel pressure chamber and by substantial financial and
other help from the Admiralty, Messrs John Aird and Son, the late
Mr Basil Ellis, Messrs S. Pearson and Son, Ltd., and Messrs Price
and Reeves.
The formation of gas bubbles in the living body during or shortly
after decompression evidently depends on the fact that the partial
pressure of the gas or gases dissolved in the blood and tissues is in
excess of the external pressure. But it is a well-known fact that
liquids, and especially albuminous liquids such as blood, will hold gas
for long periods in a state of supersaturation, provided the super-
saturation does not exceed a certain limit. In order to decompress
safely it is evidently necessary to prevent this limit being exceeded
before the end of decompression. Whether or not the decompression
is free from risk will depend on the degree of supersaturation which
can be borne with safety, the extent to which the blood and tissues
have had time or opportunity to become saturated, and the extent
to which they have had time to become desaturated again during
decompression. In carrying out our investigations we have kept
these three factors constantly in view, and it is necessary to discuss
them in some detail before proceeding further.
1
The Eeport of this Committee, which has recently appeared as a blue-book, contains
a full account of the experimental investigations on Diving, carried out under its auspices
at Portsmouth, ofi the West Coast of Scotland, and elsewhere, during the last two years :
also a short summary of the experiments detailed in the present paper, and many data as
to the occurrence of compressed-air illness in connection with diving and other work in
compressed air. The conclusions and recommendations of the Committee are summarised
at the beginning of the Beport.
PART I. THEORETICAL.
A. The rate of saturation of the body with nitrogen during

exposure to compressed air.
When a roan or animal is placed in compressed air, the blood
passing through the lungs will undoubtedly take up in simple solution
an amount of gas which will be increased above normal in proportion
to the increase in partial pressure of each gas present in the alveolar
air. The experiments of Haldane and Priestley1, which have since
been extended by Hill and Greenwood2, show that the partial pressure
of CO2 in the alveolar air remains constant with a rise of atmospheric
pressure: hence there can be no increase in the amount of CO2 present
in the blood during exposure to compressed air. As regards oxygen,
the amount in simple solution in the arterial blood will certainly
increase in proportion to the rise in alveolar oxygen pressure ; but
as soon as the blood reaches the tissues this extra dissolved oxygen,
which (except with exposures to enormous pressures) is only a small
part of the total available oxygen in the arterial blood, will be used up,
so that in the tissues and venous blood there will be at most only a
very slight increase in the partial pressure of oxygen. For practical
purposes therefore we need only take into consideration the saturation
of the body with nitrogen.
In view of what is known as to the ease and completeness with
which the blood becomes aerated in its passage through the lungs,
there seems no reason to doubt that in compressed air the blood
reaching the lung capillaries must become instantly saturated with
nitrogen at the partial pressure existing in the alveolar air (see p. 351).
At the commencement of .exposure to compressed air this blood, on
being carried to the tissues, will by diffusion share with them its
excess of nitrogen and then return to the lungs for a fresh charge.
By the constant repetition of this process the tissues, and the venous
blood leaving them, will gradually become more and more saturated
with nitrogen at the partial pressure of the nitrogen in the alveolar
air, which will be practically the same as in the inspired air. Since
the rate of blood supply and the solubility of nitrogen per unit mass of
tissue vary greatly in different parts of the body, the rate of saturation
1
Journal of Physiology, vol. xxxn. (1905), p. 229.
2
Proc. Roy. Soc, B, vol. LXXVII. p. 442.
will vary correspondingly. We may however form some rough general
idea of the average rate of saturation by assuming as a basis of calcula-
tion that the blood is evenly distributed throughout the body, and that
the tissues are similarly constituted in all parts.
According to the figures adopted by Bohr1, 100 c.c. of blood take
up in simple solution at the body temperature 087 c.c. of nitrogen
for each atmosphere of air pressure. This is only 8 °/o less than
would be taken up by water under the same conditions. Blood
contains nearly the same percentage of solids as the semi-liquid tissues
(apart from fat) in most parts of the body, and we may assume that
these tissues will take up nearly the same proportion of nitrogen as
blood. The earthy constituents of bone (about 3 °/0 of the body weight)
probably take up no nitrogen. On the other hand the body fat, as was
recently shown by Vernon2, who made a number of determinations at
the body temperature with special reference to our investigations, takes
up about six times as much nitrogen as an equal weight of blood. The
body of a well-nourished man probably contains fully 15 "/ii of its weight
as fat or fatty material. Hence it may be estimated that it will, when
saturated at any given pressure, on an average take up, weight for
weight, about 7O°/o more nitrogen in simple solution than the blood
under the same conditions, and that the whole body of a man weighing
70 kilos will take up about one litre of nitrogen for each atmosphere of
excess pressure.
Now the weight of the blood in man is about 4-9 °/o of the body
weight3: hence the amount of nitrogen held in solution in the body,
when it is completely saturated at any given pressure, will be about
170
T q , or 35 times as great as the amount present in tbe blood alone.
If therefore the blood distributed itself evenly and at the same rate
throughout the body, the latter would have received, at the end of one
complete round of the blood after sudden exposure to high pressure of
air, one thirty-fifth of the excess of nitrogen corresponding to complete
saturation. The second round of the circulation would add one thirty-
fifth of the remaining deficit in saturation, i.e. 1/35 x 34/35 of the total
excess : the third round would add 1/35 x (34/35 x 34/35) of the total
excess, and so on. On following out this calculation, it will be found
that half the total excess of nitrogen would have entered the body
1
Nagel's Handbuch der Physiologie, vol. i., 1905, p. 63.
2
Proc. Roy. Soc, vol. LXXIX. B, 1907, p. 366.
3
Haldane and Lorrain Smith, Journal of Physiology, vol. xxv., 1900, p. 340.
A. E. BOYCOTT, G. C. 0. DAMANT AND J. S. HALDANB 347
after 23 rounds of the circulation, three-fourths after 46 rounds, seven-

eighths after 69 rounds, and so on. The progress of the saturation of
the body with nitrogen is thus a logarithmic curve of the form shown in
Figure I1.
. —' "
80
70
/
BO
/
SO
/
30
I
20
10
( 2 3 4 5
Multiples of the time required to produce half-saturation.
Fig. 1. Curve showing the progress of saturation of any part of the body with nitrogen
after any given sudden rise of air pressure. The percentage saturation can be read
off on the curve, provided the duration of exposure to the pressure, and the time
required to produce half-saturation of the part in question, are both known. Thus
a part which half-saturates in one hour would, as shown on the curve, be 30°/0
saturated in half-an-hour, or 94 % saturated in 4 hours.
Experiments on animals have shown that the venous blood entering

the lungs contains about two-fifths less of oxygen than the arterial
blood. If we assume that the same proportion holds good for a man
at rest, and that very little oxygen is used up in the lungs themselves,
the percentage of oxygen gained by the blood in the lungs must be
about 8°/o> o r about double the percentage diminution in the expired
1
This calculation is in principle similar to that made by Zuntz (Fortschritte der
Medizin, 1897, No. 16), and worked out more fully by Heller, Mager and v. Schrotter
(loc. cit.). On account, however, of the discovery that fat has a very high coefficient of
absorption for nitrogen, and that the blood volume in man is considerably less than was
formerly supposed, our calculation gives a much slower rate of saturation per round of the
circulation.
air as compared with the inspired air. The volume of blood passing
through the lungs is therefore about double the volume of air breathed.
Since this volume of air (measured dry and at standard pressure and
temperature) averages about seven litres per minute1 for a man of
70 kilos during rest, the volume of blood passing through the lungs
may be estimated at about 3*5 litres per minute2. The total blood
volume is however also about 3'5 litres, so that a volume of blood equal
to the total blood volume probably passes through the lungs about
once a minute during rest. We may therefore substitute minutes for
rounds of the circulation in the above calculation of the rate of satura-
tion of the body with nitrogen, so that, if the assumptions made for the
purposes of the calculation held good for a man exposed to compressed
air, his body would be half saturated with the excess of nitrogen in
23 minutes, three-fourths saturated in 46 minutes, etc.
In reality, however, this calculation affords at best only a very
rough general idea of the actual rate of saturation, since it is known
that the distribution of blood per unit of body weight through various
parts of the body varies greatly, and that the rate of circulation
through any given part varies according as the part is at rest or in
a state of activity. The proportion of fat and fatty material is also
very different in different parts of the body, so that the capacities of
different tissues for taking up nitrogen must vary accordingly. We
should expect therefore that some parts of the body will saturate much
more rapidly than the calculation shows, and other parts much more
slowly. Direct experimental evidence of far more rapid saturation in
some parts of the body has recently been furnished by Hill and
Greenwood3. Their method was to determine the free nitrogen in
samples of urine secreted shortly after exposure to high pressure, and
shortly after return to normal pressure. A sufficiently copious secretion
of urine was produced by previously administering large drinks of water
to the subject of the experiment; and they found that, within about
ten minutes of exposure to high pressure, samples of urine secreted
were saturated at this pressure. Conversely, on lowering the pressure
to normal, the excess of nitrogen disappeared within a few minutes.
1
Haldane and Priestley, loc. cit., p. 245.
2
As a result of numerous experiments on man with the lung catheter Loewy and
v. Schrotter (Untersuchungen ilber die Blutcirculation beim Menschen, 1905, p. 90) infer
that the average rate of blood flow during rest is slightly faster. At present, however,
there is some doubt as to the interpretation of results obtained by the lung catheter
method.
3
Proc. Roy. Soc. B, vol. LXXIX., p. 21, 1907.
These results seem to show conclusively that the kidney substance

became saturated with nitrogen at a rate about ten times as great
as would correspond to the above calculation. From the data given
it appears, however, that urine was being secreted with great rapidity
during the experiments. For instance, 135 c.c. were secreted in five
minutes in one observation where the quantities and times are recorded.
This is about thirty times the average rate of secretion, so that the
circulation of blood th rough the actively working kidneys must have
been greatly increased.
Equally clear evidence of the existence of a far slower rate of
saturation is afforded by the experience of men working in compressed
air, particularly in caissons and tunnels at moderate pressures. It is
well known to those practically familiar with such work that the risk
of symptoms occurring on decompression depends on the duration of
the exposure. There is very little risk on rapid decompression after
short exposures of less than an hour to an excess pressure of two
atmospheres or even somewhat higher pressure; but as the duration
of exposure increases hour by hour, so do the risks on decompression
increase. We are assured by Mr E. W. Moir (of the firm of Messrs
S. Pearson and Son, Ltd., Westminster), who has had an exceptionally
large experience of tunnelling work in compressed air at excess
pressures up to about 2£ atmospheres, that the maximum of risk is
not reached after even three hours, so that a limitation of working
shifts to three hours markedly diminishes the frequency of compressed-
air illness. Hence in some parts of the body saturation with nitrogen
must still be incomplete after three hours. Another observation
pointing in the same direction is that when the daily working period
was 8£ hours under pressure with two intervals of about l j hours each
for meals at ordinary atmospheric pressure, cases of caisson disease
usually occurred after the last decompression in the evening and not
when the men came out for meals1.
Our own observations on animals afford fresh evidence bearing in the
same direction. We found that in goats the risks on decompression
increase with the length of exposure to pressure up to from two to three
hours (see below, p. 396).
In different warm-blooded animals the rate of respiratory exchange
varies, roughly speaking, according to the ratio of body surface to
weight. The smaller the animals, therefore, the greater is the respira-
tory exchange per unit of body weight, and the more rapid must be the
1
G. W. M. Boycott, Trans. lust, of Civil Engineers, vol. CLXV., 1906.
circulation. In consequence small animals, when placed in compressed
air, must saturate their tissues more rapidly in proportion to their more
active respiratory exchange ; and, conversely, they will free themselves
more rapidly, during or after decompression, from the excess of nitrogen.
Hence results obtained with small animals as to the time required for
complete saturation, or for safe decompression, are not directly appli-
cable to man. We selected goats for our experiments as they were the
largest animals which could be conveniently used; but their weights
averaged only about one-fourth to one-third of the weight of an adult
man. As the surfaces of different mammals are roughly as the cube
roots squared of their weights, we should expect that in goats of this
size the respiratory exchange per kilo of body weight would be about
two-thirds greater than in man. Direct determinations showed that
this was the case (see p. 381). Hence if it required three hours exposure
to a high pressure to effect practically complete saturation1 of the more
slowly saturating tissu es of a goat with nitrogen, about five hours would
be required for a man. An inspection of Fig. 1 (p. 347) will show that if
these tissues became 50 °/o saturated in about 45 minutes in goats and 75
minutes in man, they would be 94 °/o saturated in three hours for goats,
and in five hours for man. A higher degree of saturation than this
would scarcely be appreciable, and we have concluded that for practical
purposes any slower rate of saturation than this, and correspondingly
slower rate of desaturation, need not be allowed for, unless the percentage
of fat in the body is abnormally high. We must admit, however, that
there is some evidence, both from our own experiments and from practical
experience in work in compressed air, that in the parts of the body
which are the seat of " bends" a still slower rate of saturation may
exist.
B. The rate of desaturation of the body with nitrogen during

and after decompression.
If the pressure is rapidly diminished to normal after exposure to
saturation in compressed air, and no gas bubbles are liberated in the
body, it is evident that for each part of the body the curve of desatura-
tion will be similar to that of saturation, provided the physiological
conditions are constant. The venous blood will give off practically the
1
The only method apparently available to determine the time of complete saturation
in normal animals is to subject them to a series of experiments in which the pressure and
decompression are kept constant and the time of exposure varied, and to observe when the
effects cease to become any worse. The method is open to obvious limitations.
whole of its excess of dissolved nitrogen during its passage through the
lungs1, and at each round of the circulation will bring back a fresh
charge of nitrogen (at the partial pressure existing in the tissues) to
be given off. The parts which become half desaturated by this process
in a given time will be three-fourths desaturated in double the time, and
so on. The slowest saturating tissues will thus, in accordance with our
previous calculation, take one and a quarter hours to become half
desaturated in man.
The normal combined gas pressure of nitrogen, oxygen and CO2 in
the tissues and venous blood may be estimated as about 90 °/o of an
atmosphere, so that if the nitrogen pressure be more than an eighth
above normal the total gas pressure will be above atmospheric pressure.
Supposing therefore that before decompression the most slowly satu-
rating parts of the body (i.e. those half saturating in one and a quarter
hours) had been saturated to an excess pressure of two atmospheres of
air, it would take about five hours at atmospheric pressure to reduce
this excess pressure to a sixteenth (or an eighth of one atmosphere) and
so bring down the total gas pressure in the parts in question to about
atmospheric pressure. The slowness of desaturation must be as clearly
borne in mind as the slowness of saturation, in connection with all the
phenomena of compressed-air illness.
If gas bubbles are formed in consequence of too rapid decompression,
they will naturally tend to increase in size by diffusion into them, in
whatever part of the body they may be except the arteries, for some
time after the end of decompression. They may thus easily cause
blocking of small vessels, and even if they are carried to the right side
of the heart or the pulmonary arteries, and lodge there, they will
increase in bulk until the total gas pressure in the mixed venous blood
falls to one atmosphere. The same remark applies to bubbles which
1
In view of the enormous surface (probably more than 100 square metres) presented
by the lung alveoli for diffusion it seems hardly possible to doubt that the blood during its
passage through the lungs becomes saturated or desaturated to almost exactly the pressure
of nitrogen in the alveolar air. According to the calculations of Loewy and Zuntz (Die
physiologischen Grundlagen der Sauerstoff-Therapie in Michaelis' Die Sauerstqfftherapie,
Berlin, 1904), a difference in partial pressure of oxygen of less than 1 mm. of mercury
would account for the diffusion of 250 c.c. of oxygen per minute through the alveolar walls.
With a difference in partial pressure of nitrogen of two atmospheres, or 1520 mm. of
mercury, between the blood and the alveolar air only about 70 e.c. of nitrogen would
require to pass per minute in order to establish complete saturation, or desaturation, of
the blood. The conditions are thus enormously more favourable for the taking up or
giving off of this nitrogen than for the taking up of oxygen by diffusion during normal
respiration.
lodge in the branches of the portal veins. If small bubbles are carried
through the lung capillaries and pass, for instance, to a slowly desatu-
rating part of the spinal cord, they will there increase in size and
may produce serious blockage of the circulation or direct mechanical
damage. Apart from this increase of size the air bubbles passing along
the arteries are probably too small to cause any harm. Once formed
they will under ordinary conditions take a long time to become re-
absorbed, since even after the gas pressure in the blood and tissues has
fallen to normal, the excess of nitrogen pressure in the bubbles over
that in the blood and tissues will only be about a tenth of an
atmosphere at most. In one case we found bubbles in the veins of
an animal which died two days after suffering from severe decompres-
sion symptoms (see below p. 421).
In order to avoid the risk of bubbles being formed on decompres-
sion, it has hitherto been recommended that decompression should be
slow and at as nearly a uniform rate throughout as possible. We must
therefore carefully consider the process of desaturation of the body
during slow and uniform decompression. For convenience in calculation
we may imagine the process as occurring in a series of time-intervals,
the first half of each of which is spent at the pressure existing at the
beginning of the interval, and the second half at the pressure existing
at the end. Let us suppose, for instance, that the body has been
completely saturated with nitrogen at an excess pressure of five
atmospheres of air, and that decompression occurs at a rate of one
atmosphere in 20 minutes. The process may be divided into five
periods of 20 minutes, during each of which the pressure falls one
atmosphere. We can then easily calculate how far desaturation will
have gone at the end of each period, and from these data construct
a desaturation curve.
Let us first consider the mean desaturation rate of the whole body,
assuming that, when the pressure is suddenly raised or diminished to a
certain level, the tissues will on an average saturate or desaturate
themselves by 50 "/„ in 23 minutes, which was shown above to be
a probable average rate. A reference to the curve (Fig. 1) shows that
ten minutes' exposure to the reduced pressure of four atmospheres in
excess will reduce the saturation by 28 °/o of the difference between five
and four atmospheres, i.e. by 0'28 of an atmosphere. Hence at the end
of 20 minutes the tissues will on an average be saturated to 4'72 atmo-
spheres. Ten more minutes at four atmospheres will reduce the
saturation to 4'5 atmospheres, and ten minutes at three atmospheres
A. E. BOYCOTT, G. C. C. UAMANT AND J. S. HALDANE 353
will further reduce it by 28% of 4-5-3, i.e. by 042 atmosphere.

Hence at the end of the second twenty minutes the saturation of the
tissues will be 408 atmospheres. Continuing this calculation we get
the desaturation curve shown in Fig. 2, from which it will be seen that
when atmospheric pressure is reached the tissues are still saturated to
an excess pressure corresponding to 1*4 atmospheres of air.
Fig. 2 also shows a similar curve for the parts which saturate and
desaturate most slowly, and which, according to our previous calcula-
tions, take one and a quarter hours to become half saturated. At the
end of decompression these slowly desaturating parts, as shown on the
curve, are still saturated to 315 atmospheres. This of course represents
a most formidable excess ; and, as will be shown below (p. 401), uniform
decompression at this rate is dangerous even to goats, and would
certainly be extremely dangerous to men, who desaturate a good deal
more slowly than goats.
5 • -
_
— _
*
*""-•
n
10 20 30 40 50 60 70 BO 100
Time in minutes.
Fig. 2. Desaturation during uniform decompression after complete saturation at
5 atmospheres excess pressure. The thick line represents the air pressure : the
upper and lower thin lines represent respectively the progress of desaturation in
parts of the body which half saturate in 75 and 23 minutes.
Inspection of Fig. 2 shows that with uniform decompression the

nitrogen pressure in the body lags behind that of the air, and that (in
the case of the slowly desaturating parts) the amount of the lag
increases during the whole time of a decompression lasting 100 minutes.
No other result seems possible, and actual experiments point strongly
in the same direction, as will be shown presently. We must em-
phatically dissent from the conclusion drawn by Heller, Mager and
v. Schrotter that decompression at the uniform rate of 20 minutes an
atmosphere prevents any dangerous retention of gas in the body. To
prevent a maximum lag of more than one atmosphere, it would be
necessary to decompress at a rate of over one and a half hours an
atmosphere if the decompression were uniform and from an excess
pressure of five atmospheres1.
The examples given will be sufficient to illustrate the extreme
slowness with which desaturation must occur with a uniform rate of
decompression. This slowness has never hitherto been recognised, but
must evidently be reckoned with in devising measures for the prevention
of caisson disease.
It is clear that the rate of desaturation might be hastened by either
(1) increasing the difference in nitrogen pressure between the venous
blood and the air in the lungs, or (2) increasing the rate of blood
circulation. In either case the blood would give off through the lungs
an increased amount of the excess of nitrogen in a given time.
In order to increase the difference in nitrogen pressure between the
venous blood and the alveolar air it has been proposed to give a diver
oxygen to breathe during, or before decompression. As long, however,
as the pressure was above about one atmosphere in excess, or 15 lbs., it
would be impossible to do this safely, since, as will be explained more
fully below, the effects might be rapidly fatal owing to oxygen poisoning-
The possible applications of oxygen are thus somewhat limited, while
the complications involved would be very considerable. The same end
can, however, be attained in another way, as will be shown in the
following section.
The rate of blood circulation can be increased considerably by
muscular exertion. Quite moderate exertion is sufficient to increase
the respiratory exchange to three or four times the normal; and the
rate of blood flow through the lungs must be increased to something
approaching to a corresponding extent. Unfortunately, the increased
blood flow is chiefly through the muscles which are working, but
probably many parts of the body participate to a greater or less extent
in the extra blood supply. Muscular work must correspondingly increase
the rate of saturation of the body with nitrogen. For this reason it
seems desirable that where work has been done in compressed air, so
that the muscles and associated tissues have probably become rapidly
saturated with nitrogen, there should also be muscular exertion
1
It is evidently a mistake to assume that a given rate of uniform decompression, such
as 20 minutes per atmosphere, is either necessary for safety in all cases, or would be
actually safe except from some limit of pressnre. From a pressure below this limit the
rate will be unnecessarily slow, and from above it dangerously fast.
A. E. BOYCOTT, G. C. C. DAMANT AND J. S. HALDANB 355
during decompression. The rate of desaturation will thus be increased

so as to compensate for the increased rate of saturation. In the case of
short exposures to compressed air, as in diving work, this is specially
important. Even, however, when there has been no special muscular
work in the compressed air movements of joints and massage of the
skin etc. will probably hasten desaturation. This has been clearly
pointed out by Hill and Greenwood1.
Another method which can be employed for increasing the circula-
tion in the case of divers is to restrict the air supply, so that the partial
pressure of CO2 in the air of the helmet may rise sufficiently to stimulate
the respiration and circulation. Both methods are now used in the
Royal Navy during the ascent of divers.
C. The limits of safety in decompression.

It is a fact well known to those practically acquainted with work in
compressed air that even with very rapid decompression there is no risk
of caisson disease unless the pressure has exceeded a certain amount.
It seems perfectly clear that no symptoms occur with less than one
atmosphere2 of excess pressure, however long the exposure may be.
Whether any distinct symptoms ever occur with less than about
T25 atmospheres (18^ lbs. per square inch or 41 feet of sea water) seems
very doubtful: at any rate they are very exceptional. At pressures
a little above 1"25 atmospheres occasional slight cases begin to be
observed, and their frequency and gravity rapidly increase with higher
pressures unless the time of exposure is limited or slow decompression
is resorted to. The lowest pressure at which we have been able to find
any record of a death occurring from caisson disease is 23 lbs. or
1*6 atmospheres3. As will be seen below, we were able to obtain slight
symptoms on rapid decompression in 1 out of 22 goats after long
1
Proc. Roy. Soc. B, vol. LXXVII., p. 449, 1906.
a
One atmosphere or 760 mm. of mercury = 14-7 lbs. per square inch, about 1 kilogram
per square centimetre, 34 feet of fresh water, 33 feet of sea water. In this paper where
pressures are defined in pounds or atmospheres without qualification, reference is intended
to the excess over atmospheric pressure as shown on gauges, not to the absolute total
pressure.
3
Babington and Cuthbert, Dublin Quarterly Journal of Medical Sciences, vol. xxxvi.,
1863, p. 312. In the list of fatal cases given by Heller, Mager and v. Schrotter (Luft-
druckerkrankungen, p. 1072), are entered two deaths at a pressure of 1-4 atmospheres.
A perusal of Paul Bert's original account (La Pression Sarometrique, p. 401) shows that
both the pressure and the cause of death are quite uncertain.
356356 The
ThePrevention
Preventionof ofCompressed-air Illness
Compressed-air Illness
exposure
exposure (four(four hours)
hours) to to136136 atmospheres
atmospheres or or 20 20 lbs.lbs.With With 25 25 lbs.lbs.
(1*7(1*7
atmospheres) two cases of slight illness occurred
atmospheres) two cases of slight illness occurred out of 23 animals. out of 23 animals.
If the risks
If the risksof ofrapidrapiddecompression
decompression depended
depended simply
simply on on thethe extent
extent
to towhich the blood and tissues are supersaturated
which the blood and tissues are supersaturated with nitrogen with nitrogen on on
decompression, we should expect to find
decompression, we should expect to find that even a short exposurethat even a short exposure to to
such an excess pressure as two atmospheres
such an excess pressure as two atmospheres would be risky with rapid would be risky with rapid
decompression
decompression : for there
: for cancan
there be be no no doubt doubt thatthatwithin,
within, say,say,halfhalfan an hour hour
or forty minutes the tissues, and the blood
or forty minutes the tissues, and the blood returning from them, must returning from them, must
be befor for
all all
practical
practical purposes
purposes fullyfullysaturated
saturated in inmany many partspartsof of thethe body,body,
andand particularly in parts of great physiological
particularly in parts of great physiological importance which importance which areare
richly
richlysupplied
supplied with withblood.
blood.Nevertheless
Nevertheless it seems
it seems to be
to bewellwell established
established
thatthata man
a man may stay without serious risk for forty minutes aatpressure
may stay without serious risk for forty minutes at a pressure
which
which would
would involve
involve greatgreatdanger
danger on onrapid rapiddecompression
decompression if he remained
if he remained
in it
in for several
it for several hours.
hours.
Parts
Partsof of thethebody bodywith witha arapid rapidcirculation
circulationwillwillbecome becomeveryvery
completely saturated in a comparatively
completely saturated in a comparatively short time, but the short time, but the highly
highly
supersaturated blood which first returns
supersaturated blood which first returns from them on rapid decom- from them on rapid decom-
pression
pression cancan remain
remain butbuta verya very short short time time supersaturated
supersaturated duringduring each each
round of the circulation, and on reaching
round of the circulation, and on reaching the large veins will mix the large veins will mix with with
lessless
highly
highly saturated
saturated blood
blood from from otherother parts parts of of thethe body.body. It would It would
seemseemthatthat
thethe state of high
state of high supersaturation
supersaturation in anyin any portion
portion of ofblood bloodlastslasts
for for
tootooshort a time to enable bubbles
short a time to enable bubbles to form. to form.
If this interpretation
If this interpretation of the
of the facts is correct,
facts is correct, we we should
should expectexpect to find
to find
withwith small animals, which rapidly saturate and desaturate, thathigher
small animals, which rapidly saturate and desaturate, that a a higher
pressure
pressure would
would be be required
required to toproduce
produce symptoms
symptoms on onrapidrapid decompres-
decompres-
sionsionafter
after a long exposure than in the case of larger animals.TheThe
a long exposure than in the case of larger animals.
general
general experience
experience of ofprevious
previous observers
observers is in is inaccord
accord with with this, andand
this, ourour
ownown experiments (see below p. 402) showed
experiments (see below p. 402) showed that we could produce that we could produce no no
obvious
obvious effects
effects in in mice,
mice,andand very very fewfew in in rabbits,
rabbits, rats,rats,andand guinea-pigs,
guinea-pigs,
by by sudden decompression after exposures
sudden decompression after exposures at pressures which at pressures which were were
invariably or frequently fatal
invariably or frequently fatal to goats. to goats.
SinceSincesupersaturation
supersaturationto to thetheextent extentof of about aboutT25T25atmospheres atmospheres
above normal atmospheric pressure
above normal atmospheric pressure can be borne with can be borne with impunity,
impunity, though
though
a greater degree of supersaturation is
a greater degree of supersaturation is risky, it seems clear that,risky, it seems clear that, in in
decompressing after prolonged exposure to
decompressing after prolonged exposure to high pressures, the rate of high pressures, the rate of
decompression
decompression should
should be besufficiently
sufficiently slow slowto to prevent
prevent anyany greater
greater excess
excess
of of
saturation than this in any part of the
saturation than this in any part of the body at the end of decom- body at the end of decom-
pression.
pression.OnOn thethe otherother hand hand decompression
decompressionshould should evidently
evidently be be as as
rapid as is possible, consistently with safety.
rapid as is possible, consistently with safety. A pressure of 1 to 1"25 A pressure of 1 to 1"25
atmospheres
atmospheres aboveabove normal
normal corresponds
corresponds to to from from2 to 2 to 2'252'25 timestimes thethe
normal atmospheric pressure; but the volume (not the mass) of gas
(measured at the existing pressure) which would be liberated if the
whole excess of gas present in supersaturation were given off is the
same whether the absolute pressure is reduced from two to one
atmospheres, or from four to two, or from eight to four. Hence it
seemed probable that, if it is safe to decompress suddenly from two
atmospheres of absolute pressure to one, it would be equally safe to
decompress from four atmospheres absolute to two, from six atmo-
spheres absolute to three, etc. Our experiments, which are detailed
below (p. 398), have shown that this is the case1. The process of
desaturation can therefore be hastened very greatly by rapidly reducing
the absolute pressure to half, and so arranging the rest of the decom-
pression that the saturation in no part of the body shall ever be
allowed to correspond to more than about double the air pressure.
The main advantage of this plan is that the discharge of nitrogen from
the tissues is from the outset of decompression increased to the greatest
rate which is safe. The rate of discharge evidently depends on the
difference in partial pressure of nitrogen between the venous blood and
the alveolar air; and by keeping this difference at the maximum
consistent with safety a great saving of time is effected. Detailed
investigations have completely justified the adoption of this principle :
they are described below, and comprise, besides a series of observations
on animals, a number of experiments in which Lieut. Damant and
Mr Catto were exposed to excess pressures up to 80 pounds, or 6"4
atmospheres of absolute pressure, in the experimental chamber and to
93^ pounds, or 74 atmospheres, in actual diving. The method greatly
simplifies the problem of safe decompression, and gets rid of many
practical difficulties, particularly in connection with deep diving. It
may be conveniently referred to as the method of " stage decom-
pression," and is so described in the sequel, though its essential
peculiarity does not lie in the decompression being done in stages
but in its being rapid till the absolute pressure is halved and slow
afterwards.
1
Whether the law holds good for pressures much exceeding six atmospheres is still
doubtful, as no experimental data exist.
Journ. of Hyg. vm 23
D. Practical measures for avoiding Compressed-air Illness.
From the foregoing discussion the general nature of the measures

needed to prevent compressed-air illness will be evident enough. The
risks may best be avoided by properly calculated stage decompression, or
by cutting down the period of exposure to a safe limit, or by both methods
combined. In the case of work in compressed air in caissons, tunnels,
etc., it is for economic reasons very undesirable to greatly reduce the
period of exposure. In diving work, on the other hand, the periods of
exposure are generally short in any case, and they can, without great
inconvenience, be confined within limits which largely reduce the risks
of compressed-air illness. Long periods of decompression are also very
undesirable in diving, since changes of weather or tide or other causes
may render a return to the surface necessary without any long delay in
coming up, and since very prolonged stays under water are exhausting,
and the diver's hands may become benumbed by cold.
As our investigations were in the first instance made with the
object of securing safety from compressed-air illness in diving work, we
may first consider the precautions desirable in connection with diving.
(1) Diving work.

The ordinary diving dress (Plate IV) consists of a copper helmet
screwed to a corselet, the latter being in its turn connected water-tight
to a stout water-proof dress covering every part of the body except the
hands, which project through elastic cuffs. Air is supplied through a
non-return valve on the helmet from a flexible pipe connected with an
air-pump on a boat or ship. The air escapes through an adjustable
spring valve at the side of the helmet. The arrangement is thus such
that the pressure of the helmet air breathed by the diver is always at
least equal to, and usually slightly greater than, the pressure of the
water at the valve outlet. At a depth of 33 feet or 10 metres the
diver is therefore breathing air at an excess pressure of one atmosphere,
or at an absolute pressure of two atmospheres; and every additional
33 feet will add another atmosphere to the pressure. To enable the
diver to sink, the dress and boots are suitably weighted. He is usually
in connection with surface by a life-line containing a telephone wire,
as well as by the air-pipe.
JOURNAL OF
JOURNAL OFHYGIENE,
HYGIENE, VOL.
VOL. VIII.
VIII. NO.
NO. 33 PLATE IV
Diving dress,
Diving dress, front
front view,
view, with
with air-pipe
air-pipe and
and life-line,
life-line, which
which are
are
connected with
connected with the
the helmet
helmet behind.
behind.
In descending or ascending a diver usually makes use of a rope

attached to a heavy sinker at the bottom. He can thus easily regulate
the rate of his ascent or descent, and take care that this rate is not so
rapid as to cause any discomfort or pain in the ears owing to incomplete
opening of the Eustachian tubes. A too rapid descent or ascent might
cause mechanical injury followed by middle ear inflammation.
As explained above, there appears to be practically no risk of
symptoms occurring from liberation of gas bubbles on rapid decompres-
sion if the pressure has not exceeded 1"25 atmospheres, corresponding
to a depth of about seven fathoms or 42 feet of sea water. Up to this
depth therefore no special precautions against caisson disease need be
taken1. At greater depths precautions depending on the duration
of exposure are evidently needed. The precautions which we have
calculated to be desirable are embodied in the table given below
(Appendix IV.); and the principles and experimental results on which
this table is based must now be discussed.
It will be convenient to consider first the case of diving to a very
great depth, and we shall take as an extreme example the case of
exposure at a depth of 35-J- fathoms (213 feet) of sea water, corre-
sponding to an excess pressure of nearly 6'5 atmospheres, or an absolute
pressure of 7'5 atmospheres.
Let us first suppose that the body of a diver is completely saturated
with the nitrogen of air at this pressure, and that it is required to
conduct his ascent to surface as rapidly as possible but without any
risk of symptoms due to bubble formation, i.e. in such a way that, in
accordance with the principles already laid down, the nitrogen pressure
in no part, of the body shall ever be more than double that of the air
breathed at the same time.
The first step would obviously be to reduce the absolute pressure to
about half, i.e. from 7"5 atmospheres absolute to 3*75 or from 6"5
atmospheres in excess to 275. This would be ex hypothesi the greatest
initial drop in pressure which would be perfectly safe. The remainder
of the decompression would evidently need to be conducted in such a
way that the maximum partial pressure of nitrogen in any part of the
body should diminish at double the rate of the fall in absolute pressure
of the air. The ascent of a diver can be conveniently regulated from
1
Heller, Mager and v. Schrotter recommend that at all depths decompression should
be at a rate of at least 20 minutes per atmosphere. This would imply a delay of 25 minutes
in coming up from 42 feet. Both common practical experience and our own experiments
show that this excess of caution is quite unnecessary at small depths.
23—2
the surface by signalling to him to stop or come on at every ten feet as
indicated on the pressure gauge attached to the pump. We may
therefore divide the ascent into stages of ten feet, and the short periods
occupied in the actual ascents may be neglected.
Since the depth was 213 feet, corresponding to 246 feet of water in
absolute pressure, it would be safe to come up at once to a depth
corresponding to 123 feet of absolute water pressure, i.e. to 90 feet of
actual depth. Consequently the first stage would be a rapid ascent of
123 feet, and it would be necessary to wait here before the next ascent
of 10 feet until the maximum partial pressure of nitrogen in the body
had fallen to that of the nitrogen in air at 2 x (80 + 33) = 226 feet of
absolute water pressure. The difference between 246 and 226 is 20,
and this is 16°/0 of 213 — 90 = 123, the difference between the original
and the reduced pressure. The most slowly desaturating parts of the
body will, according to our previous calculations, take 75 minutes to
give off half of any excess of nitrogen which they may contain at any
given air pressure; by inspection of the curve (Fig. 1) it will be seen
that they will take about 19 minutes to lose 16°/o °f the excess.
Hence a delay of 19 minutes would be necessary at 90 feet before
coming up to 80 feet. At 80 feet the partial pressure in the body
would require to fall an amount corresponding to 20 feet, which is
about 17-J- % of 193 — 80 = 113, the new difference in relative pressure
between the nitrogen in the body and in the air. This would necessi-
tate a delay of 21 minutes before ascending to 70 feet. The further
delays needed would be 23 minutes at 70 feet, 26 minutes at 60 feet,
30 minutes at 50 feet, 35 minutes at 40 feet, 42 minutes at 30 feet,
51 minutes at 20 feet, and 62 minutes at 10 feet. It would thus take
309 minutes, or more than five hours, to reach surface.
This calculation is represented graphically in Fig. 3. It will be
noticed from the figure that the time required for safe decompression
does not increase proportionally to the increase in depth. For instance,
an increase in depth of 15 feet from 50 to 63 feet necessitates an
increase of 45 minutes in the time required for safe decompression;
but the same increase in depth from 198 to 213 feet only requires an
increase of 15 minutes in the time of decompression.
A somewhat more rapid rate of stage decompression could probably
be adopted without appreciable risk to life, but the occurrence under
water of even one of the less serious decompression symptoms might be
extremely unpleasant or indirectly dangerous, so that a factor which
we believe to be thoroughly safe in this respect has been used in the
calculation. The possible occurrence of slight symptoms after surface

had been reached would not, however, be a serious matter: for this
reason half of the last stop at 10 feet from surface might be dispensed
with, which would save half an hour. The most slowly desaturating
tissues would, according to the calculation, still be only saturated to an
excess pressure of 13 atmospheres—a safe enough limit perhaps, but
leaving no great margin to spare.
Fig. 3 also shows the maximum excess of saturation with uniform
decompression in the same time and in 10 hours. It will be seen that
uniform decompression in about five hours would leave at the end of
decompression an excess saturation within the body of 2'1 atmospheres;
and even if uniform decompression were extended to ten hours the
excess saturation would still exceed one atmosphere. It is also per-
fectly clear that uniform decompression is an unsuitable way of
bringing a man out of compressed air. Where a sufficiently safe rate
of uniform decompression is employed (as, for instance, with 10 hours
in the case under consideration), it is only at the very end (when the
nitrogen pressure inside the body becomes more than double that of
the air) that there is any risk of symptoms occurring; and for the
sake of safety at the end the whole process is made quite unnecessarily
long. Increased safety at the end is only secured in combination with
useless delay at the beginning1.
As will be seen in Part II, the results of our experiments, allowance
being made for the difference between goats and men, fully confirm the
foregoing mode of calculation. Not only has stage decompression in
the calculated time proved safe where uniform decompression in the
same total time was unsafe, but shorter periods of stage decompression
than those calculated have been proved to involve risk of symptoms,
increasing in gravity and frequency with the shortening of the time,
though always less than the risk from uniform decompression in the
same time.
If the whole body of a diver were allowed to become saturated at
any great depth, it is evident that the time needed for safe decom-
pression would be impracticably long. To reduce the time of de-
1
The regulations of the Dutch Government make the following method of decompres-
sion obligatory for work in caissons, &c. The pressure is to be lowered at the rate of not
more than ^ t h of an atmosphere in 3 minutes till 3 atmospheres of excess pressure is
reached: then at not more than T^th of an atmosphere in 2 minutes till 1J atmospheres
excess pressure is reached; and finally at not more than TVth of an atmosphere in 1J
minutes till normal pressure is reached. This method is still more unsuitable than
uniform decompression, and would be very unsafe with high pressures.
compression to within limits practicable for divers, it is evidently
necessary to greatly reduce the period of exposure to high pressure1.
At great depths limitation of the exposure is also necessary in order to
avoid toxic effects from the high pressure of oxygen (see p. 371).
Calculation of the mode and period of decompression required after a
limited exposure to pressure is a somewliat complicated matter, but the
principles already laid down render it quite possible.
2 3
Time in hoars.
Fig. 3. Theoretical ascents of a diver after a prolonged stay at 213 feet of sea water.
Stage decompression in 309 minutes compared with uniform decompressions in
309 minutes and in 10 hours. Continuous lines = stage decompression : interrupted
lines = uniform decompression. Thick lines = air pressure: thin lines = saturation
with atmospheric nitrogen in parts of the body which half saturate in 75 minutes.
When a diver goes down for a very short time, we have to take into
consideration not only the time which he spends at the maximum
pressure on the bottom but also the time occupied in the descent and
the ascent. During the descent he is all the time saturating himself
with nitrogen, and during most of the ascent he may be doing so also.
Calculation will show that, if he descends and ascends at a uniform
rate, the time spent in this process will be nearly equivalent, as regards
the saturation of the body with nitrogen, to half the same time spent
at the maximum depth. It is therefore clear that in deep diving the
diver should descend as rapidly as is practicable, and should also
ascend at once, on completion of his work, as far as he safely can.
The rate of descent may be limited either by pain in the ears or by an
air supply insufficient to keep the upper part of the dress full of air.
1
This was fully realised by Catsaras who recommended a stay on the bottom of only
1 minute at 30 fathoms.
A. E.
A. E. BOYCOTT,
BOYCOTT, G.
G. C.
C. C.
C. DAMANT
DAMANT AND
AND J.
J. S.
S. HALDANE
HALDANE 363
363
Both these
Both these causes
causes are
are avoidable,
avoidable, andand an
an experienced
experienced diver,
diver, with
with his
his
Eustachian tubes well opened and a proper supply of air,
Eustachian tubes well opened and a proper supply of air, can get to ancan get to an
excess pressure of six atmospheres (198 feet) in two
excess pressure of six atmospheres (198 feet) in two minutes. This minutes. This
time was found
time was found sufficient
sufficient in
in experimental
experimental dives
dives up
up to
to 210
210 feet
feet made
made byby
Lieut. Damant
Lieut. Damant andand MrMr Catto
Catto (Appendix
(Appendix II).II). TheThe recommendation
recommendation
commonly made
commonly made that
that the
the rate
rate of
of both
both ascent
ascent and
and descent
descent should
should bebe
slow is evidently quite unsound. A man who spent
slow is evidently quite unsound. A man who spent half an hour in half an hour in
descending to 30 fathoms, and an equal time in ascending
descending to 30 fathoms, and an equal time in ascending at a uniform at a uniform
rate, would
rate, would runrun aa considerable
considerable risk
risk of
of perishing
perishing on on his
his return
return toto the
the
surface.
surface.
tO
tO 20
20 30
30 so
so eo
eo 70
70 80
80 90
90 WO
WO 120
120 130 HO
130 HO
Time in
Time in minutes.
minutes.
Fig. 4.
Fig. 4. Desaturation
Desaturation during
during stage
stage decompression
decompressioninin32
32minutes
minutesandanduniform
uniformdecompression
decompression
in 22 hours,
in hours, after
after exposure
exposure for
for 15
15 minutes
minutes at
at 75
75 lbs.
lbs. pressure
pressure with
with compression
compression in
in
66 minutes.
minutes. Thick Thick lines
lines ==air
air pressure:
pressure: continuous
continuous lines
lines==stage
stage decompression:
decompression:
dotted lines
dotted lines == uniform
uniform decompression.
decompression. TheThe curves
curves from
from above
above downwards
downwards represent
represent
respectively the
respectively the variations
variations in in saturation
saturation with
with nitrogen
nitrogen of of parts
parts of
of the
the body
bodywhich
which
half saturate
half saturate in in 5,
5, 10,
10, 20,
20, 40,
40, and
and 75
75minutes.
minutes.
In order
In order to
to illustrate
illustrate the
the method
method by by which
which we we have
have calculated
calculated safesafe
modes of ascent in the minimum period of time
modes of ascent in the minimum period of time we may take as an we may take as an
example the case
case of
of exposure
exposure forfor 15
15 minutes
minutes to to aa pressure
pressure ofof 75
75 pounds
pounds
(6"1 atmospheres
atmospheres absolute
absolute or or 2828 fathoms
fathoms == 168 168 feet).
feet). Many
Many of of our
our
goats were
experiments on goats were made
made with
with this
this pressure
pressure and and exposure.
exposure. ItIt
minutes to
took about six minutes to raise
raise the
the pressure
pressure in in the
the experimental
experimental
pounds, so
chamber to 75 pounds, so that
that the
the total
total virtual
virtual exposure
exposure till
till decom-
decom-
pression began was was about
about 1818 minutes.
minutes. Fig. Fig. 44 shows
shows graphically
graphically the the
variations of pressure
pressure during
during this
this period:
period: also
also thethe calculated
calculated partial
partial
pressure of nitrogen
nitrogen in in different
different parts
parts of
of the
the body,
body, asas compared
compared with with
the nitrogen pressure
pressure in in the
the air.
air. The
The first
first stage
stage waswas from
from 6"16"1 to
to 2'8
2'8
atmospheres absolute (corresponding to an ascent in sea water from
168 to 60 feet) and occupied four minutes. The subsequent stoppages
were:—
2 minutes at 2'8 atmospheres (60 feet of sea water),
3 „ 2-2 „ (40 „ ),
5 „ 1-9 „ (30 „ ),
7 „ 1-6 „ (20 „ ),
10 „ 13 „ (10 „ ).
It will be seen from the figure that this rate of decompression was
slightly faster than what was calculated above to be desirable. At
the end of decompression the nitrogen pressure in those parts of the
body which became half saturated in about 20 minutes under pressure
would be equivalent to that of air at about 1'4 atmospheres, or 20'6
pounds per square inch. If the circulation in one of these parts were
less vigorous during decompression than during exposure to the high
pressure, it might well be that the nitrogen pressure in this part at
the end of decompression would be higher than corresponded to the
calculation. As a matter of fact minor symptoms (" bends") were
observed five times in 34 decompressions of 18 goats, although no
serious effects occurred. We concluded that the period of virtual
exposure (18 minutes) was slightly longer than is desirable with stage
decompression in 31 minutes: in the table below (p. 442) the limit has
been set down at 15 minutes.
Fig. 5 shows the calculated nitrogen pressure in different parts of
the body during uniform decompression in 31 minutes after the same
exposure at 75 pounds. It will be noticed that at the end of decom-
pression there is a dangerous excess of saturation in all parts of the
body except those which half saturate in less than about seven or eight
minutes, and that this supersaturation corresponds to an excess pressure
of as much as 2 - l atmospheres of air. The goats used for the stage
decompression experiments were on alternate occasions subjected to
uniform decompression in the same time and with the same exposure.
The result was that, in 36 decompressions, one died, two were paralysed,
one had indefinite general symptoms of a severe character, and in 11
other cases " bends " occurred, besides two doubtful cases. This was
entirely in accord with what the calculation would lead us to expect;
and uniform decompression in 31 minutes is evidently dangerous under
the conditions given.
It might be supposed that safety would be secured by extending to
a moderate degree the length of uniform decompression. It must be

remembered however that the more the duration of uniform decom-
pression is extended, the longer is the period during which the body
is exposed to high pressure. Fig. 4 shows the calculated effects of
uniform decompression extended to two hours. Although the quickly
saturating parts of the body are desaturating during the greater part
of the decompression, the slowly saturating parts are, on the other
hand, becoming more and more saturated, so that at the end of de-
compression the parts which half saturate in from 40 to 75 minutes are
saturated to an excess pressure of about 1*7 atmospheres, although at
the beginning of decompression they vvere only saturated to from 07
to T3 atmospheres and could consequently have given no trouble.
to so 30 40 50
Time in minutes.
Fig. 5. Desaturation during uniform decompression in 32 minutes after exposure for
15 minutes at 75 lbs. pressure with compression in 6 minutes. Thick line = air
pressure. The curves from above downwards represent respectively the variations in
saturation with nitrogen of parts of the body which half saturate in 5, 10, 20, 40 and
75 minutes.
Very prolonged uniform decompressions are extremely tedious, and it

seemed scarcely worth while to make any extensive series of such
experiments. We found however that out of 12 goats uniformly
decompressed in 90 minutes after 18 minutes virtual exposure at 75
pounds (61 atmospheres of absolute pressure) three developed symp-
toms of bends after decompression. The proportion of illnesses was
thus greater than with stage decompression in a third of the time.
With men the results would certainly be much worse, and we
calculate that for a man, after the same exposure, several hours would
be needed for uniform decompression in order to escape all risk of
symptoms occurring. The time would in fact require to be nearly
as long as if the body had been completely saturated at the maximum
pressure.
With very short exposures to high pressure, rapid decompression is
probably safer than uniform decompression at a moderate rate. There
is a considerable human experience on this point: divers working at
great depths would seem to consider it fairly safe to go rapidly to the
bottom at a depth of 160 or 180 feet and return equally rapidly,
provided the time spent on the bottom does not exceed six or eight
minutes and provided also that the dives are not repeated at short
intervals. It is reported of the skilled Greek divers of the Mediter-
ranean that, in case their gear becomes entangled on the bottom, they
will cut their air-pipe and line and blow themselves up to the surface
in less than a minute from a depth of 30 fathoms or the like rather
than stop more than about ten minutes on the bottom. Our experi-
ments on goats are in accordance with this practice. We found that
no symptoms were produced by sudden decompression in less than a
minute after virtual exposures at 75 pounds up to four minutes, and
even in some trials up to six minutes (see below, p. 394).
With exposures exceeding a very few minutes, or such brief ex-
posures frequently repeated, so that during the intervals the body has
not time to become desaturated, we have little doubt that slow and
uniform decompression—the slower the better—is at any rate preferable
to sudden decompression. Uniform decompression must however be ex-
tremely slow to make it entirely free from risk of death or very serious
symptoms, and the time required is so great that this method seems to
us quite impracticable in connection with diving work. There appears
to be very little human experience of slow uniform decompression.
Divers usually come up in a few minutes at most, and even half an
hour spent in the ascent would appear to be quite exceptional. Almost
the only definite observations are those of Hill and Greenwood, who
recently experimented on themselves at very high pressures. Fig. 6
shows the variations of pressure and the calculated saturations of
different parts of the body during the experiment in which Greenwood
went to a pressure of 91 pounds (7"1 atmospheres absolute). This
experiment appears to have been a very risky one. After decom-
pression he had bends in both arms, and Hill also had symptoms
pointing towards blockage of vessels in the subcutaneous fat after
similar experiences at 75 pounds pressure.
In Appendix IV two tables are given for the safe decompression of
A. E. BOYCOTT, G. C. 0. DAMANT AND J. S. HALDANB 367
divers after exposure for varying periods of time at different depths.

These tables are the same as are now in use for divers in the Royal
Navy, on the recommendation of the Committee on Deep Diving. In
Table I the period of virtual exposure is so limited that the diver can
return to surface by stages in half an hour or less. It will be noted
that the maximum periods of exposure are from the time of leaving
surface, so that there should be no chance of increased danger from
undue delay in descending. The stoppages during the ascent are
so calculated that, until surface is nearly reached, the excess of
nitrogen pressure in any part of the body should never be more
than double the nitrogen pressure of the air breathed, and not more
/
\
/
/ ^
OS
/ /
a3 \
5 2 \
<j 0
Time in hours.
Fig. 6. Showing calculated variations in saturation with nitrogen during Dr Greenwood's
experiment on himself. Thick line = air pressure: the two curves from above
downwards represent respectively the variations in saturation with nitrogen of parts
of the body which half-saturate in 40 and 75 minutes.
than two and a quarter times this pressure when surface is reached. The
only case in which these limits are allowed to be slightly exceeded is
with short exposures in comparatively shallow water. This slight
excess is, however, only in parts of the body which saturate and de-
saturate very rapidly, and, as already explained, give rise to no danger.
As an additional safeguard the diver is directed to keep his arms and
legs constantly moving during each stoppage, so as to increase the rate
of circulation and guard against the chance of the rate of desaturation
during his ascent being proportionally less than the rate of saturation
during his stay on the bottom while he was doing work.
The second table provides for the case of exceptionally long stays
under water. A diver may be delayed by his air-pipe or life-line being
fouled, or by other exceptional circumstances, against which it is
necessary to provide. Where the fouling has been complicated by the
action of tide the delay on the bottom has occasionally amounted to
several hours, until the tide has slackened or turned. If the diver is at
a great depth the calculated time required for safe decompression after
so prolonged a stay is very long. On the other hand the dangers
from cold and exhaustion have to be considered, and the difficulties
caused by a strong tide during the diver's ascent. In view of these
difficulties the time allowed for decompression after very prolonged
exposures is somewhat curtailed, but not so much as to permit of risk
of more serious symptoms than " bends," in so far as experiments on
animals, and human experience, render it possible to calculate. In the
case of men of exceptionally heavy build, and inclined to obesity, the
time allowed after very prolonged exposures ought to be increased by
about a third, although such men, particularly if over about 45 years of
age, ought not to expose themselves to the risk of a prolonged stay in
very deep water.
It might appear as if the rate of stage decompression recommended
after prolonged exposures was slower than is actually required. A very
unfortunate accident which occurred recently has shown only too clearly
that this is not the case. In connection with the work of raising
a torpedo boat which had sunk in 25 fathoms (150 feet or 46 metres)
several divers were employed. They were working in 20 minute spells,
and returning to surface by stages in 32 minutes, in accordance with
the first table, which was the only one then in use. No symptoms of
any kind were observed after the divers' return to surface under these
conditions, nor have any symptoms ever been observed hitherto among
divers working according to the table. One of the divers, however,
became fouled in a very exceptional manner. His life-line was fixed in
one direction over a spar or rope belonging to the sunk vessel, and his
air-pipe was fixed in the other direction. He was thus prevented from
going to free either his air-pipe or life-line. A second diver at once
went down, but was unable to free him owing to the drag caused by the
tide; and it was only after two and a half hours, when the tide had
slackened, that he got free. He was then brought up by stages under
the direction of Staff Surgeon Rees and Lieut. Damant. For the
decompression two and a half hours were allowed, which we then
believed would be a sufficient time in case of a diver being badly fouled
at 25 fathoms. The diver was, however, a man of heavy build with

much fat in his body, and aged 49. Owing to his exhausted condition
he did not come up on the ordinary rope, but had to be pulled up,
hanging motionless on the life-line during the long stoppages. On
reaching surface he was very exhausted and could hardly have been
safely kept longer in the water. He had no paralysis or other definite
symptoms of caisson disease. A bed was arranged for him on deck, and
hot bottles &c. applied. After a time he complained of some pain in
the legs, but this soon subsided; and as he seemed much better in the
morning he was removed to hospital, since there was no suitable accom-
modation for him on the gun-boat where he was. The moving made
him worse, and he gradually became restless and delirious in spite of
administration of oxygen at intervals, showed signs of cardiac failure,
and died somewhat suddenly about 24 hours after he had been brought
up. At the post-mortem examination 12 hours later a moderate
number of bubbles were found in the right side of the heart, the veins
of the liver and intestines, while scattered bubbles were present in
vessels elsewhere, including the coronary vessels, though none were seen
in the vessels of the brain. The mesenteric fat, which was very
abundant, was in places distended with small bubbles. There was
about an inch of subcutaneous fat over the trunk, but no bubbles were
seen in this layer. There seemed no reason to doubt that death was
largely due to the bubbles, although the more usual symptoms of
caisson disease were absent. There were no signs of pneumonia.
This is the only known case of prolonged exposure of a man to such
a high excess pressure as four and a half atmospheres; and although
his age, heavy build, and exhausted condition combined to make the
circumstances very unfavourable, the fact of his death shows that the
long decompression periods recommended in the second table after
prolonged exposures are none too long, even for a man of ordinary
build. Every precaution should be taken to guard against such long
exposures at high pressures.
A diver has often to descend twice or oftener at short intervals. At
the beginning of the second descent the more slowly desaturating parts
of the body will not have had time to lose their excess of nitrogen, and
consequently they will be more highly saturated at the end of the
second descent than would otherwise have been the case. This will be
clear from a study of Figs. 2 or 3. To meet the increased risk in
decompression it is desirable, in calculating the proper stoppages, to add
together the two periods of exposure, and adopt the corresponding rate
370 The Prevention of Gompressed-air Illness
of decompression shown in the tables. For the first half of the
I
stoppages this is not necessary, but for the second half, including the
longer stoppages needed to meet the case of the more slowly desatu-
rating parts, the rule should be carried out. The increasing danger
after successive short dives by pearl divers, &c. without any precautions
in decompression is notorious. This danger does not mount up to the
same extent with stage decompression, but nevertheless exists. As
the interval between successive dives increases the added danger on
decompression diminishes. With an hour's interval the extra pre-
cautions might be halved, and with two or three hours' interval they
might be omitted.
It may be remarked that the precautions recommended in the tables
are greatly in excess of those which have hitherto been commonly
employed in either diving work, or work in caissons, tunnels, &c. We
have endeavoured to leave a clear margin beyond everything which
either human experience or experiments on animals, or calculation, has
shown to be risky. In connection with diving, the practice hitherto
recommended in the British and other navies has been that the diver
should both descend and ascend at a uniform slow rate. By abolishing
the slow descent and ascent, and substituting stage decompression, it
has been possible to combine greater safety with a clear saving of
time under water for a given working period on the bottom. Where
the air supply to the diver is managed in accordance with the recom-
mendations of the Diving Committee there is also very greatly increased
working efficiency in deep water. For a discussion of the air supply to
divers, and many other practical points relating to diving, we must refer
to the Committee's Report; and to the " Diving Manual," which has
just been re-written and issued to the Royal Navy.
A possible complication to which we have not hitherto referred in
connection with compressed-air illness arises from the fact that at very
high pressures of air the partial pressure of oxygen begins to be so high
as to be capable of producing serious effects. Paul Bert discovered that
oxygen at a partial pressure exceeding about three atmospheres
(corresponding to 143 atmospheres of air) causes animals to go into
convulsions and die, even a short exposure being often fatal. More
recently, Lorrain Smith, who experimented on mice, and whose results
have been confirmed and extended by Hill and Macleod, showed that
oxygen at high pressure acts on the lungs, producing pneumonia1. He
1
Lorrain Smith, Journ. of Physiology, vol. xxiv., p. 19, 1899; Hill and Macleod,
Jonrn. of Hygiene, vol. m., p. 401, 1903.
found that fatal pneumonia may be produced after four days' exposure to
an oxygen pressure of as little as 75 °/o of an atmosphere, corresponding
to air at an absolute pressure of 3"6 atmospheres (88 feet of sea water).
At a pressure of about 125 atmospheres of oxygen (6 atmospheres of
air, or 165 feet of water) death, from pneumonia was produced in about
48 hours. At about 1*8 atmospheres of oxygen (eight and a half atmo-
spheres of air, or 250 feet of water), marked symptoms usually occurred
in about 12 hours, and death in 20 hours, though in one case death
followed in seven hours. At about 2"8 atmospheres of oxygen (13"3
atmospheres of air, or 40(5 feet of water) marked symptoms were observed
in about three hours, and death in nine hours.
The steel chamber at the Lister Institute was not made to withstand
such high pressures as would produce within a short time symptoms of
oxygen poisoning if air alone was pumped into the chamber. We have,
however, made a few observations in the chamber when the oxygen
pressure of the air breathed was raised by other means. In one experi-
ment seven goats were placed in the chamber, and the oxygen pressure
raised by opening three large cylinders of oxygen, and at the same time
pumping in air to 81 pounds pressure. The total oxygen pressure was
thus raised to 2'3 atmospheres, corresponding to a depth of 55 fathoms,
or 330 feet, or 100 metres. After three hours one animal had died of
pneumonia in the chamber, and most of the others seemed more or less
affected, though they rapidly recovered on decompression1. We also
tried on ourselves the effects of breathing nearly pure oxygen from a
bag while we were in the chamber at an absolute pressure of two atmo-
spheres ; but we could not detect any effects after a few minutes with
an oxygen pressure of 1"7 atmospheres, corresponding to about 40
fathoms (240 feet or 73 metres). In a number of goats which were
exposed to 75 pounds' pressure (168 feet or 51 metres of water) for
three hours, no symptoms indicative of oxygen poisoning were ob-
served.
To judge from these data there is no immediate risk to a diver from
oxygen poisoning at depths up to 40, or perhaps 50 fathoms (73 to 90
metres) if ordinary air is breathed, provided the stay is not long. With
stage decompression the diver could rapidly return to a perfectly safe
oxygen pressure; but, as already remarked, we do not yet know with
1
One animal showed bends after decompression which was effected in 133 minutes by
stages. After exposure at + 75 lbs. for 3 hours in air this decompression gave 2 bends in
14 goats. There is therefore no evidence that the exposure to high pressure oxygen in-
creased the susceptibility to caisson disease.
certainty whether it is perfectly safe to rapidly reduce the pressure to
half after exposure to such very high air pressures.
(2) Work in caissons, tunnels, and diving bells.

In connection with various kinds of engineering work under water,
or in soft water-bearing strata, compressed air is commonly used for
keeping water out of the working place and preventing collapses.
The men have thus to work continuously in compressed air.
In tunnels or 'tubes' through soft water-bearing strata, where a
steel lining has to be erected to keep water out and resist pressure, the
working face, or blind end of the tunnel under construction, is kept free
of water by the air pressure with the help of a circular shield with a
cutting edge which is advanced as each section of steel lining is erected
into position. The soil is excavated by hand labour, and passed out on
trucks through an air-lock.
In constructing foundations for the piers of bridges over rivers,
caissons are employed. A caisson is a steel tube, which ultimately
forms the lining of the pier, and is shaped accordingly. Near the lower
end there is a steel diaphragm, forming a working chamber. An inner
steel tube passes through this diaphragm, and serves for ingress and
egress, and for passing up the material excavated. At the top of this
inner tube there are air-locks for allowing the passage of men and
material without escape of the compressed air contained in the working
chamber. The latter is kept free from water by the air pressure, and
the excess of air escapes beneath the cutting edge of the caisson. When
a secure foundation for the pier has been reached this chamber is
rilled up with concrete. In constructing mine shafts through soft
water-bearing strata the same principle may be employed. For work of
a simpler kind on river or harbour bottoms diving bells are often used,
the bell being simply lowered to the bottom at any required place, so
that the men can work on the area covered by it and are kept dry by
the air pressure.
The circumstances connected with work in compressed air in caissons,
tunnels, &c, differ in certain respects from those associated with diving
work.
In the first place the duration of exposure is far longer. A caisson
or tunnel worker is usually in compressed air for six or eight hours
daily, or even longer. The conditions of the work render any great
limitation of the periods of exposure very difficult and expensive.
A. E. BOYCOTT, G. C. 0. DAMANT AND J. 8. HALDANE 373
Usually, however, the workman comes out for meals at intervals of about
three hours.
A second difference is that the very high pressures to which a diver
may have to go are not needed in caisson or tunnel work. An excess
pressure of about 3{ atmospheres, or 48 lbs., is, we believe, the extreme
limit hitherto employed; and usually the excess pressure does not exceed
about two atmospheres or 30 lbs. Decompression seems to be usually
effected in 10 to 20 minutes, or even, with the lower pressures, in three
to five minutes.
With properly arranged air-locks for men and material there should
be no need for hurry in coming out; and undue. hurry is specially
undesirable if the workman leaves the works at once, since he would be
liable to develop symptoms when he was so far away that he could not
be readily recompressed. To obviate this risk as far as possible, it is
customary to endeavour to keep men for half to one hour on the
works after they come out; and with the usual rates of uniform
decompression this precaution is very necessary. Evidently, how-
ever, it is greatly preferable to prevent all practical risks of serious
symptoms.
In order to attain this end stage decompression as recommended for
divers in the tables in Appendix IV may be employed. An accurate
and easily read pressure gauge, visible from both inside and outside the
air-lock, is of course essential; and a reliable man should be in charge of
the tap. As a further control it would be desirable to have an automatic
graphic record of the variations of pressure each time the lock for men
is used. As any very sudden drop in pressure might cause mechanical
injury, the outlet tap should be so arranged as to prevent decompression
at a maximum initial rate of more than about one pound in five
seconds1. With this arrangement and an ordinary tap, the rate of
decompression would diminish considerably as the pressure fell, and
the proper point for interrupting the decompression could be accurately
reached.
The tables in Appendix IV have been calculated with special regard
to the comparatively short periods of exposure to pressure in diving work;
1
The delivery of the inlet tap should also be restricted, and the man in charge should
have strict directions to take care that the rate of admission or discharge of air does not
cause pain in the ears, &c. of any of the men in the lock. To avoid pain a very slow rate
of air admission may sometimes be needed, but with practice a rise of pressure of one atmo-
sphere per minute is often not too much, so that any definite rule, limiting the rate to
much less than this, seems scarcely desirable.
Journ. of Hyg. VIII 24
and the stoppages recommended during the divers' ascent after
exceptionally long periods of exposure are somewhat shorter than
would be desirable apart from the risks entailed by the long stay under
water. In the case of caisson and tunnel workers, on the other hand, it
is only in exceptional cases that the exposure to pressure lasts less than
three hours; and usually the exposure during the day lasts at least six
hours.
With such long exposures and only moderate pressures the calculated
theoretical rate of safe decompression after the first rapid stage is nearly
uniform; and the rules for decompression may be greatly simplified
by adopting uniform slow decompression or uniform stages1.
The following table shows the rate of uniform slow decompression
calculated to be safe after the initial diminution of absolute pressure in
the proportion of 2 : 1. Suppose, for instance, that men were working
at a pressure of 24 pounds in 3-hour spells, with an hour's interval
between for a meal. In coming out they would be rapidly decompressed
24 +15
to an absolute pressure of — s = 19-5 pounds or 45 pounds of
excess pressure. After the first 3-hour spell of work the slow decompres-
sion would be at the rate of one pound in three minutes, or 3 x 4"5 = 13^
minutes in all. After the second spell the rate would be one pound
in five minutes, corresponding to 11\ minutes in all. If they stayed for
the whole period in the compressed air the rate of slow decompression
would be one pound in seven minutes corresponding to 3 1 | minutes in
all. To take another example, if the work were at 40 pounds excess
40 +15
pressure the men could be rapidly decompressed to j. = 27|
pounds of absolute pressure, or 12J pounds excess pressure. After a first
3-hour spell of work the period of slow decompression would therefore
be 12J x 7 = 87 minutes: after a second spell (with an interval of 30 or 45
minutes outside the lock) \1\ x 8 = 100 minutes; and after a con-
tinuous exposure of six or seven hours, 12^x9 = 112 minutes2.
1
With the lock air-tight, and no ventilation, uniform decompression at any required
rate could be easily secured by means of a reducing valve on an outlet, with a graduated
tap beyond it, the arrangement being similar to the reducing valve and tap usually
connected to a cylinder of compressed oxygen or gas used for limelight. If the delay in
the lock is so long that ventilation is required, or if ventilation is needed in order to
compensate for accidental leakage, it would be best to have an adjustable safety valve on
the outlet, and adjust this by one pound at a time at the proper intervals.
2
We have some doubt as to whether the increased slowness of decompression after
very long exposures would be altogether sufficient to meet the increased tendency to slight
symptoms (" bends "). These are, however, of minor importance if all serious symptoms
TABLE I.
Table showing rate of decompression in caisson and tunnel work.
Number of minutes for each pound of decompression
after the first rapid stage
r
After second or After six hours
Working pressure After first third three hours' or more of
in pounds per three hours' exposure, following an continuous
square inch exposure interval for a meal exposure
18—20 pounds 2 3 5
21—24 ,, 3 5 7
25—29 „ 5 7 8
30—34 „ 6 7 9
35—39 „ 7 8 9
40—45 ,, 7 8 9
It will be evident from the last example that in order to avoid waste
of time in the lock it would be preferable with pressures exceeding
about 25 pounds to keep the men under pressure continuously during
each shift. Thus with two 3-hour spells of work separated by a
decompression, the time spent in the lock would be 87 + 100 = 187
minutes; whereas if the meal were taken in the compressed air, the
two 3-hour spells would only imply 112 minutes in the lock.
With working pressures exceeding about 25 pounds the air-lock
should be roomy and comfortably arranged, and large enough to take
the whole of a shift of men. It should be provided with an electric
heater, telephone, and if possible some sort of lavatory accommodation.
With pressures up to 45 pounds, or four atmospheres of absolute
pressure, there appears to be no substantial objection to keeping men for
six hours, or even more, continuously under pressure, provided that the
mode of decompression is thoroughly safe. With pressures exceeding
about 40 pounds, the practice has hitherto been to limit, the exposure to
about one hour, and employ rates of decompression which are danger-
ously rapid. This plan implies greatly increased risk and expense, since
for the accomplishment of the work the number of decompressions is six
times as great, and the men are idle most of the day. The actual
increase in risk must be very great.
In tunnel work, or any other kind of work where plenty of space is
available, there would be great advantage in providing a large air-lock,
or section of tunnel, in which the pressure was constantly maintained at
a little less than half the absolute pressure in the working section.
are prevented. We also think that with long shifts, exceeding a total of about 3 hours,
still slower decompression would be needed for any men inclined to obesity. Such men
should, therefore, be excluded in the medical examination which all men working in air at
high pressures ought previously to undergo.
24—2
The men could then pass rapidly (in two or three minutes) from the
working section into this intermediate lock or section, where they could
take their meals, wash, and change their clothes. After a sufficient delay
(dependent on the working pressure) they could then pass out rabidly.
If, for instance, the working section was at a pressure of 30 pounds, the
intermediate or " purgatory" lock could be kept at an absolute pressure
of about —^r— = 20'5 lbs., or 5£ pounds of excess pressure '. At the
end of the day's work there would be a delay of about 50 minutes
in this large lock, during which the men could wash and change, or take
a meal. With this plan all delays during actual decompression would
be obviated, so that ingress and egress would be free at all times, and
the men could use the locks employed for material. For persons going
in for only short periods the delay in the "purgatory" lock could
be curtailed in accordance with the tables in Appendix IV. The move-
ment of the men while employed in washing, changing clothes, &c. would
hasten the process of desaturation, and this would be a further
advantage.
In any case where it was specially desirable to reduce the period of
delay in the air-lock to a minimum, recourse could of course be had to
breathing oxygen during the period of slow decompression. This would
about double the rate of desaturation, and therefore halve the delay.
The oxygen could be breathed from a bag, and the CO2 absorbed by a
purifier, so that very little oxygen would be needed. By so arranging
the mouthpiece that part of the expired CO2 was rebreathed, and the
respiration and circulation thus stimulated, a still better result would be
attained.
The results of some of our experiments seem to indicate that even
the very slow rate of stage decompression which has been recommended
above would be insufficient to completely obviate the risk of "bends"
occurring after prolonged exposure. The rate of saturation and
desaturation of some of the tissues which are the seat of "bends" is
possibly slower than we have provisionally assumed. What we
have aimed at is to completely obviate the risk of any serious
symptom, while at the same time reducing the chances of "bends"
to a minimum.
1
A comparatively rapid fall in absolute pressure in the proportion of 2-2 to 1 is within
practically safe limits, particularly if the previous period of continued exposure has not
exceeded three or four hours.
o
cz
33
z
>
r-
o
JTI
p
CO
The steel chamber at the Lister Institute. View from outside, The steel chamber at the Lister Institute. Front end,
showing the back end of the chamber, with the large door showing the manhole for entering, the small air-lock for
and one inspection window. passing food, &c. into the chamber, an inspection window, -o
I—
a pressure gauge, and several valves, &c.
PART II. EXPERIMENTAL.
1. Apparatus.
We owe the large pressure chamber (Plate V) in which both
human and animal experiments were conducted to the generosity of
Dr Ludvvig Mond, F.R.S. It is a short segment of a boiler of
§ inch plate resting on its side ; the ends are slightly dished steel
plates £ inch thick. Inside it measures 7 | feet long by 7 feet wide and
high, and has a capacity of 9500 litres (336 cubic feet). It is thus large
enough to hold 3 or 4 persons comfortable and can be used for
animal experiments lasting several hours without the necessity of
ventilating. There are two doors : one, an oval manhole (24 x 15 inches),
is easily removed and is in common use; at the other end is a large
rectangular plate (28 x 24 inches) which can be unbolted for the
admission of bulky articles. There are a number of spring and simple
valves; the largest is in the floor of the chamber and serves also as
a drain; when fully opened it reduces the pressure from 100 lbs. to
atmospheric pressure in rather less than a minute. Besides this there
are four spring and three simple valves so arranged that the pressure
can be completely controlled either from inside or outside. The front
is also furnished with an air-lock, by means of which small articles
can be passed in or out of the chamber during an experiment.
Three windows are provided of stout glass; as a precaution for safety
these are fitted with an arrangement whereby the breaking of the glass
releases a solid metal rubber-faced plug which falls into the hole.
Wiring for lights, a telephone, electric heaters and a motor to drive
a fan, kymograph, &c, is introduced through fibre plugs.
The pressure is raised or reduced by a simple compressor driven by
a gas engine. While this has proved quite satisfactory for negative
pressure experiments, the rate at which the pressure can be raised by its
means is only about 2 lbs. per minute. This was a serious obstacle to
the examination of the effects of exposure to high pressures of short
duration. Accordingly after the preliminary experiments, a multitubular
compressed air reservoir was placed at our disposal by the Admiralty.
This reservoir has a capacity of about 22 cubic feet, and by charging it
to about 70 atmospheres with a two-stage liquid-air compressor and also
another steel bottle to 180 atmospheres we were enabled to suddenly
blow the contents into the chamber and so reach a pressure of 60 lbs. in
4 minutes, and 75 lbs. in 5|—G^ minutes according to the temperature.
The pressure is indicated by two Sehaffer spring gauges (one of which is
visible from within the chamber) for positive pressures, and one spring
gauge outside and a mercurial barometer inside for negative pressures.
The spring gauges show a lag of nearly 2 lbs. up to about half an
atmosphere, but above one atmosphere they are concordant and, as
far as could be ascertained, correct.
The chamber and accessory apparatus have now been frequently used
during eighteen months for experiments at pressures varying from
100 lbs. above to 8 lbs. below atmospheric pressure, and have been found
very satisfactory and convenient.
2. Choice of earperimental animals.

A few experiments were made with rabbits, guinea-pigs, rats, and
mice, but for regular use goats were selected chiefly because they were
the largest animals which could be conveniently dealt with and which
could be obtained in considerable numbers. The questions under con-
sideration depend in a very fundamental way upon the rate of circulation
in the animal under investigation. Among the ordinary mammals this
must vary with the rate of the respiratory exchange per unit of body
weight and is therefore proportional to the ratio between body
surface and body weight. The susceptibility of any animal to caisson
disease after sufficiently long exposure to compressed air must depend
in the main upon the rate at which its respiration and circulation
removes the excess of dissolved nitrogen on decompression.
Not only is this excess removed more rapidly in small animals, so
that the time during which bubbles might be formed is correspondingly
less, but, as already pointed out, there is every reason to believe that
the time during which the venous blood remains in a supersaturated
state during each round of the circulation largely determines the forma-
tion of bubbles. This time is so short in small animals that no bubbles
at all are formed, in spite of the temporary existence of very great
supersaturation in the blood and tissues. The susceptibility of any
species of animal then varies enormously with the size. Thus a mouse,
weighing 20 grammes and with a CO2 production of about 8 grammes
per kilo per hour, is much less susceptible than a goat, weighing 20,000
grammes and producing about 0'8 gramme CO2. We have indeed
failed to produce any symptoms at all in mice on decompression in less
than a minute after one hour's exposure at 75 lbs., an experience
invariably fatal to goats. In the same way dogs, with a respiratory

exchange of some 13 gms. CO2 per kilo per hour, are much less sus-
ceptible than men with an exchange of about* 0*5 gms. Thus Heller,
Mager and v. Schrotter observed no symptoms in dogs1 on sudden
decompression from any pressure less than about 60 lbs., while abundant
illnesses are caused in man, and for the matter of that in goats also, by
inappropriate release from 30 lbs. pressure. It therefore appears clear
that it is necessary to use large animals for experiments which are
designed to illustrate the incidence of caisson disease in man. Indeed
the quantitative factor by which the results obtained on quite small
animals might be translated into human experience is so large as to
become qualitative in character.
Since pressures of some 100 lbs. or more are required to produce
symptoms in a reasonable proportion of small animals, the use of animals
such as goats is also very desirable in order to keep as far away as
possible from the point at which the partial pressure of oxygen is high
enough to cause toxic effects. We found that an exposure of three
hours at 81 lbs. to an atmosphere containing 36 °/o oxygen (the oxygen
pressure being thus equal to that of 150 lbs. excess air pressure) killed
one goat out of seven with "pneumonia." Our experience shows that
it is not necessary to exceed an air pressure of half this (75 lbs.) to pro-
duce symptoms which are sufficiently varied and severe to satisfy
experimental requirements.
Experience also showed that goats were very suitable animals in that
slight symptoms were presented to our notice in a definite objective
form. The lesser symptoms of caisson disease cannot be neglected, and
there are reasons for supposing that their occurrence is not exactly
conditioned by those experimental circumstances which in a more severe
form produce serious and fatal results. They cannot be properly
detected in mice or guinea-pigs or even in rabbits. Goats, while they
are not perhaps such delicate indicators as monkeys or dogs, and though
they are somewhat stupid and definitely insensitive to pain, are capable
of entering into emotional relationships with their surroundings, animate
and inanimate, of a kind sufficiently nice to enable those who are
familiar with them to detect slight abnormalities with a fair degree of
certainty.
The animals, 85 in number, used in the present experiments were a
mixed collection of ordinary English goats of no particular breed. They
1
The weights are only given in a few instances; from these it may be surmised that
the dogs were small (5 to 12 kilos).
were about equally distributed between the sexes, and varied in weight
from 10 to 30 kilos, the average being rather less than 20 kilogrammes.
All were apparently adult, judging from the fact that none showed
any increase in weight while in our possession. One or two (XIII A,
XXXII A) seemed to be quite aged, but the rest were fully active.
On the whole the herd remained healthy. Two died of apical
pneumonia and two of diarrhoea, which was at one time epidemic in a
severe form. The cause could not be determined, but the trouble
became much less marked after the animals were placed on a more
meagre diet and corn withheld. Three animals were under some
suspicion of being infected with M. melitensis1; two of them seemed
rather depressed (though not more so.than appears to be natural in
some goats), while the third showed no signs of ill-health. Various
items of pathological interest were found in those which came to post-
mortem : in the lungs various nematodes were found several times,
Linguatula once, and a surgical needle once; a Streptothrix abscess in
the stomach wall followed puncture with a needle to relieve distension ; a
bony tumour was found in an adrenal gland ; in one old goat (XXXII A)
the aorta was extensively, atheromatous; flukes occurred in the liver
once, while hydatids in the peritoneum were very common and intestinal
worms abundant. None of these conditions (except possibly the arterial
disease) can however be considered to have rendered the animals
definitely abnormal as far as caisson disease was concerned, and none of
them could be attributed to exposure to compressed air.
3. Respiratory exchange of goats.

The difficulties of measuring directly the circulatory activity of
normal animals are almost insuperable. This must be however in
general proportionate to the rate of respiratory exchange, and a number
of determinations of the CO2 production of our goats were made in order
to get a line of comparison with other animals (and especially man) in
respect of the rate at which air would be taken up by and discharged
from the body.
1
Of 22 animals whose blood was examined, 16 gave no reaction with M. melitensis at
a dilution of 1: 20, 3 gave some reaction at 1: 20, while 3 animals gave complete aggluti-
nation up to 1:200 (XVIII A, XXVI A, XXIII A). Cultures from the blood during life
were negative, and when they eventually came to autopsy cultures of blood, spleen, liver,
inguinal, axillary, mesenteric and mammary glands were negative as regards M. melitensis.
The exact history of these animals could not be obtained, but there is practically no doubt
that they had never been out of England.
A. E. BOYCOTT, Gr. C. C. DAMANT AND J. S. HALDANE 381
The observations were made by using the pressure-box as a respira-

tion chamber. The animals were enclosed and hourly samples removed
(after thorough mixing with an electric fan) and analysed in a delicate
form of Haldane's gas analysis apparatus. The results were entirely
satisfactory, the successive analyses showing a regular increase in the
CO2. The goats led a regular life, and all the observations were made
at approximately the same time of day so that they are fairly comparable
with one another in respect of the influence of food. The animals
remained fairly quiet, though they seldom lay down.
The results of 27 experiments are given in the next table. The
analyses have been calculated in grammes of CO2 per hour per kilo of
TABLE II.
Goats CO2gms. per hour
•sg
2*
If s
Male
S§ 88-S 3|
3 X m 5-8 £0 & s EH 5s & t*" Remarks
*° 19 764 4 4 99-0 24-7 1-006 2-625
I 13 763 3 0 4 2 2 62-8 15-7 1123 2-070
II 15-5 765 2* 0 9 4 5 159-7 17-7 0-908 1-823
III 15 752 14 0 9 3 6 1711 19-0 0-727 1-749
IV 17 762 11 0 6 6 — 111-8 18-6 1104 2-554 B. Q. 1 0 3 .
V 11 769 3 0 6 — 6 138-8 231 0-670 1-771 E. Q. 0-90.
VI 13 765 6 0 6 6 — 121-3 20-2 0-975 2-389 B. Q. 106.
VII 13 754 5 45 7 *J 142-7 20-4 0-887 2-187
VIII 13 740 4 45 8 — 8 193-7 24-2 0-627 1-682
IX 15 754 4 0 7 7 — 1421 20-3 0-664 1-630 B. Q. 0-91.
X 16 780 7 0 6 6 — 126-3 210 0615 1-533 Fasting 20 hrs.
XI 6
XII 15 778 21 7 7 — 140-9 20-1 0-763 1-770
4
XIII 15 774 0 8 — 8 193-7 24-2 0-548 1-469 B. Q. 0-82.
7
XIV 14 760 45 13 5 8 295-7 22-7 0-667 1-738 B. Q. 1-08.
4
XV 16 758 25 6 6 — 127-9 21-3 0-959 2-367
34 20 7 3 4 21-2 0-635 1-572
XVI 17 764 • 148-7
XVII 15 762 0 6 6 — .127-9 21-3 0-669 1-652
XVIII 17 762 45 6 — 6 1221 20-3 1-020 2-504
XIX 14 761 45 6 6 — 127-9 21-3 0-697 1-722
XX 13 760 0 5 — 5 100-1 20-0 0-921 2-258
XXI 16 760 45 5 — 5 1001 20-0 1104 2-701
XXII 15 762 45 6 127-9 21-3 0-967 2-390
XXIII 16 768 45 5 100-1 200 0-852 2-083
XXIV 12 776 0 4 95-3 23-8 0-717 1-853
XXV 14 775 0 4 95-3 23-8 0-751 1-941
XXVI 14 775 0 4 76-9 19-2 0-624 1-501
XXVII 13 766 0 4 76-9 19-2 0-704 1-693
body weight and also per 1000 square centimetres of surface according
to the usual formula S x 100 = ^ ^ x 11-2, where S= surface in square
centimetres and W the body weight in kilogrammes.
The goats used belonged to Series II (Exps. 1—4), III and IV.
The results of these experiments are very variable; the averages are
shown in the next table :
TABLE III.
CO2 in grms. per hour
No. of Per kilo Per 1000 sq. —\
experiments body-weight cms. surface
At atmospheric pressure 16 0-795 l"907
At 45 lbs. positive 8 0-853 2-126
At 20, 21 and 25 lbs. positive 3 0-786 1-903
All pressure experiments 11 0-834 2065
Males only 10 0-830 2-019
Females only 8 0-843 2-137
Mixed experiments 9 0-762 1-771
All experiments 27 0-811 1-971
(410 c.c.) (997 c.c.)
One may conclude that goats produce about 0-8 grm. CO2 per kilo per
hour under conditions of incomplete rest, and that no great departure from
this figure is occasioned by the animals being under pressure up to 45 lbs.
or by sex. It is shown elsewhere1 that something more than lO°/o of the
total CO2 produced by goats comes from the fermentation of the contents
of the alimentary canal, and figures detailed below (p. 409) indicate that
one-fifth of the body weight is contributed by these contents. In
comparing the CO2 production of goats with that of man, we may
regard these two corrections as roughly balancing one another and may
neglect them.
It appears that man produces under conditions of bodily activity
comparable to that of our experimental animals, about 045 to 0'5 grm.
CO2 per kilo per hour. Goats therefore show a respiratory activity
approximately 1'7 times that of man. This figure corresponds fairly well
with that calculated from the size. If the respiratory exchange per
unit of surface is the same, a goat of 20 kilos will produce 1*5 times as
much CO2 per unit of weight as a man of 70 kilos.
4. Method of conducting the experiments.

No animals were subjected to experiment when obviously ill. As
a rule five to eight animals were put in at one time. The pressure
1
Journal of Physiology, vol. xxxvi. (1907), p. 283.
A. E. BOYCOTT, Q. C. C. DAMANT AND J. S. HALDANB 383
having been raised to the desired point, the chamber was entirely
closed and no ventilation given until decompression began. The average
CO2 production of goats is about 435 c.c. per kilo per hour at ordinary
temperatures. The chamber usually contained 100 to 150 kilos of goat
so that the CO2 rose about 045 to 055 °/o (measured at atmospheric
pressure) per hour. In this way it never attained a harmful partial
pressure in experiment^ lasting from a few minutes to four hours. In
the few observations made with an exposure of eight hours, the CO2 was
allowed to accumulate for four hours and afterwards the chamber was
ventilated so that the CO2 did not exceed a partial pressure of 2 °/0 of
an atmosphere. No experiments have been made to directly examine
the possible influence of CO2 upon the incidence of caisson disease. It
appears to the authors that the effect must (1) in any case be very
slight with partial pressures of less than 2 or 3 °/o> a nd the result, if
any, of the increased respiratory and circulatory activity must be in the
direction of diminishing the ill-effects of decompression after any but
quite short exposures1.
After the preliminary experiments (Series I), the animals were never
used more than once on the same day, and, with rare exceptions, not on
succeeding days. In many cases indeed individual goats rested for a
week or more between the experiments.
During decompression the animals could be watched fairly satis-
factorily through the windows of the chamber, though fog of course
completely blocked the view during the actual moments of rapid
decompression. At the end they were allowed to escape from the
chamber and run about free in the yard. They were kept under
continuous observation for half an hour or longer, and were frequently
seen throughout the day. We found that practically all the symptoms
which were going to appear declared themselves within thirty minutes,
though a few slight signs were probably missed. We also found that
slight signs were much more obvious when the animals were not
distracted or excited by food or other causes. During the breeding
season it is advisable to keep the males and females separate, and,
by removing any sources of interest, to allow the animals to fall into
a state of meditative boredom. Under these circumstances, trivial
symptoms are easily detected which are not made the subject of
objective demonstration by animals engaged with their appetites.
1
Greenwood (British Medical Journal, June 22nd, 1907, Supplement, p. 409) has
recently found that high percentages of CO2 do not increase the liability to decompression
symptoms.
No observations were made of the temperature within the chamber
during an experiment. Very hot and very cold weather did not seem
to influence the results. The air in the chamber was always warmed by
compression and sometimes also artificially, while decompression was
of course accompanied by sudden, often very severe, spells of cold. No
account has been taken of variations in atmospheric pressure. The
extreme readings of the barometer on record are 806 and 689 mm. at sea
level, and in this country 790 and 695 mm.1, giving ranges of 117 mm.
and 95 mm. or about 2\ and 1 | pounds. Even this variation, though it
occurs at an important part of the absolute pressure scale, cannot be of
great significance.
Times of exposure of one hour or less are, unless the contrary is
directly specified, to be taken as indicating actual exposure to the given
pressure, the time of compression (six minutes) being neglected. For
longer exposures it was sometimes convenient to raise the pressure
more slowly: in these cases therefore the times specified may indicate
either the actual exposure plus four to six minutes compression or
a virtual exposure calculated by adding the actual exposure to half the
time of compression which is in minutes roughly one quarter of the
pressure in pounds positive (see above, p. 362).
As will be gathered from the details given below, the general scheme
of the experiments involved the examination of three variable factors—
degree of pressure, duration of exposure and duration and mode of
decompression. For the most part the degree of pressure was kept
constant while the other two factors were varied. It soon appeared from
the preliminary experiments that the individual variability of the
animals was very large—larger indeed than the difference between many
of the modes of decompression which it was desired to examine. It also
appeared that the relative susceptibility of the different individual
animals remained fairly constant so that after a time one could pick
out goats which were known to be either susceptible above the average
or definitely resistant to caisson disease. It was therefore clear that
either an enormous number of animals had to be employed or the
experiments had to be so framed as not to produce fatal results and so
reduce the proportion of susceptible individuals in the herd. It appears
probable that any 20 or 30 goats would give much the same results, but
if many are lost it is necessary to discard the remainder and procure a
fresh batch to be subjected to the comparative experience. Obvious
reasons prevented this procedure. It was therefore necessary to be at
1
Nature, vol. LXXV., 1907, p. 330.
JOURNAL OF HYGIENE, VOL. VIII. NO. 3 PLATE VI
' Bends " of fore-leg in a goat.

some pains to secure that the deaths should be as few as possible so

that the same individuals might pass through a number of different
combinations of pressure, exposure and di compression. The animals
were therefore first put through the experiments which we surmised
would give least symptoms, and were subsequently exposed to circum-
stances of progressively increasing severity. Even so, each batch of
animals became selected to a more or less considerable degree. Be it
noted however that one may in this way obtain strong evidence of an
a fortiori kind. For, if the selected resistant members of the herd
show many symptoms in the severest experiments, so much the more
would the whole original lot of average animals have been affected.
This individual variability of the animals renders many of our experi-
ments incomplete, and should be constantly borne in mind in considering
the results obtained.
5. The symptoms observed in goats.

The symptoms observed in goats in sequence to decompression are
protean in character. The majority may however be grouped under a
few definite heads.
1. Bends. The commonest symptom which we have observed
consists of the exhibition of signs indicating that the animal feels
uneasy in one or more of its legs. The limb, most commonly a fore-leg,
is held up prominently in the air and the animal is evidently loth to
bear weight upon it (see Plate VI1). In mild cases such a limb is used
normally in walking or running, but in other instances the animal
limps more or less considerably when it is forced to use the affected
member, and is often very anxious to lie down. No tenderness can be
detected on pressure or manipulation of the leg and it is not altogether
clear that the animal suffers definite pain. We have however noted
that a goat may break its leg and immediately use it for progression
without evincing any signs of pain. We may conclude from this that
the response to stimuli which in many animals would be distinctly
painful is largely suppressed in the goat to the level of the exhibition of
a consciousness that the limb is somewhat abnormal and not well suited
for active use. But it must be understood that this objective demon-
stration is a very conspicuous and definite symptom. There is little
doubt that these symptoms observed in the legs of goats are the
equivalent of the "bends" or "screws" which are the commonest
1
We are indebted to Dr H. W. Armit for this photograph.
symptoms in caisson workers; in human experience they are of course
accompanied by definite pain, often of a severe character.
The following table shows the distribution of "bends" in the last 110
cases observed:
TABLE IV.
One hind leg ... ... ... ... ... 28
One fore leg ... ... ... ... ... 70
Both hind legs... ... ... ... ... 1
Both fore legs ... ... ... ... ... 1
One fore and one hind leg ... ... ... 10
Total one leg ... ... ... ... ... 98
Total two legs ... ... ... ... ... 12
Total right ... ... ... ... ... 50
Total left ... ... ... ... ... 48
"Bends" maybe seen immediately at, or indeed (but very rarely)

shortly before the end of a long decompression. Most commonly
however they come on after an interval of about 15 minutes; on the
other hand they may be delayed still more. As might be expected, the
period of delay varies with the duration of decompression : thus the
average delay in a number of cases after rapid decompression (1 to 10
minutes) was 16 minutes, which was reduced by long decompression to
six minutes. Their duration appears to be brief; all evidence of
their presence has usually disappeared in one or two hours and
it has been very exceptional for any trace of them to be present next
day (16 to 20 hours).
" Bends " in parts of the body other than the limbs are very difficult
to identify in animals; we have however occasionally noted symptoms
which might well be bends in the trunk, though we are not prepared to
definitely identify them as such.
2. Temporary paralyses may be of two kinds. In the first a
general weakness is present accompanied with dyspnoea and there is
dragging of the hind legs with foot-drop. These are clearly symptoms
due to a general deficiency of oxygen from pulmonary embolism and are
comparable to the paralyses seen in, e.g., carbon monoxide poisoning in
animals and men. In our records and the tables such cases are not
classified as " paralysis " but as " dyspnoea." In the second group fall
a series of cases which are obviously of nervous origin. The animal,
while showing no signs of general illness, or in other instances having
already had bends, exhibits foot-drop or a more extensive palsy in one
or more hind- or fore-limbs. The paralysis does not usually come on
till about 15 minutes after decompression, rapidly becomes more marked
A. E. BOYCOTT, d C. C. DAMANT AND J. S. HALDANB 387
for a few minutes after the first signs are noted, and then soon
begins to mend, so that there is marked improvement in about
half an hour, and by next day the animal is found quite well. This
form of paralysis chiefly involves the hind legs (16 out of 19 cases).
3. Pain. In some cases the animals have shown signs of acute
pain by urgent bleating and continual restlessness. Bleating in goats
after decompression is usually a sign of distress such as is produced by
cardiac and respiratory embarrassment and is often present in fatal cases.
In other instances animals showing only severe bends bleat in a most
distressing manner and are evidently in acute pain: at the same time
they may gnaw at some part of their body (such as the testicles) as if
localising the origin of the pain. In animals which have recovered, we
have not had any instance where these signs persisted for more than 10
or 15 minutes.
4. Permanent paralyses. The onset is usually immediately after
decompression, the condition is complete from the first and for at
least several days there are no signs of improvement. In a few
cases the first paralysis has passed off (to all appearances com-
pletely) in two or three hours and the animal has been found next
morning to be again paralysed. This second paralysis is permanent.
A similar history has often been noted in human cases. In 15 cases out
of 16 the condition has been a paraplegia, and in one all four legs were
affected more or less. In some there has been retention of urine, and
one animal had to be killed on account of acute distension of the
stomach which came on some 20 hours after the onset of the paraplegia.
In the most severe cases the animals have been killed; others have
however soon begun to mend and have lived for some months with a
slight spastic paralysis of the hind legs.
5. A fair number of cases have occurred where the animal has
been obviously ill, but in which it has been impossible to identify any
definite local symptoms or any definite dyspnoea. The goat may lie
down, refuse to move or to be tempted with corn (of which goats are
inordinately fond), sometimes lying extended on the side, sometimes
hurriedly rising, walking a few steps and then lying down again. On
two occasions the most probable interpretation of the symptoms was
that the animal was blind. The goat may run wildly about instead of
becoming very apathetic and depressed. These and other such symptoms
are on the whole somewhat persistent and the animal is often dull and
poorly the next day. In one case (XXV A) the goat showed little but
a marked apathy and distaste for food, but died 16 hours later.
6. Dyspnoea is usually the precursor of (7) death and only a
minority of goats survived after showing clear dyspnoea. In these
cases the condition has rapidly improved; more commonly however it
progressively increases till the animal is moribund, when it is replaced
by irregular, faint, gasping respiration. The mucous membranes become
livid and pale and the animal lies for a short time unconscious before
respiration stops. The heart continues to beat regularly throughout
and the rate is not apparently much altered. Only on one occasion
have we been able to hear gurgling in the heart on auscultation: it was
then audible at some distance. Death ensues at varying periods after
decompression; with very severe experiments (e.g. 100 lbs.: 1 hour:
1 minute) l it may follow in five or ten minutes: with more moderate
conditions it is delayed for 20 or 30 minutes, or rarely for two or three
hours : on three occasions it has followed still later, up to 40 hours. The
delay in the onset of the first symptoms is often most striking; the
animal may appear quite normal for as long as 10 or 15 minutes, dyspnoea
then appears, the goat falls down helpless and in another 15 minutes
is dead.
8. Mechanical symptoms are not important. We have not been
able to satisfy ourselves that goats ever suffer materially during
compression from the ear troubles which are so common in men.
Abdominal distension is occasionally extreme, but the animal soon
empties its distended stomach and seems to be little inconvenienced2.
Our index throughout has been the presence of symptoms, not the
presence of bubbles. Anticipating here a later section (p. 410) we may
say we are in entire agreement with the view which attributes most of
the severe symptoms of caisson disease to local or general blocking of the
circulation by bubbles of gas. One might suppose in consequence that
the incidence of severe symptoms, especially of paralyses, would be of ,a
haphazard kind, since they would be to a large extent dependent on the
chance distribution of bubbles by the blood stream. Some support for
this view is perhaps to be found in the records of caisson workers given
by von Schrotter ; as far as can be ascertained from the details given, the
cases of paralysis and dyspnoea were distributed through the whole
range of pressure experienced by the men in about the same proportion -
to the total number of illnesses of all kinds, which latter increased
greatly as the pressure became higher. It should however be noted that
the range of pressure was small (up to 2'4 atmospheres positive), and
1
i.e. pressure 100 lbs. positive; exposure for 1 hour; decompression in 1 minute.
2
Post-mortem experience shows that the stomach alone is distended, not the bowels.
the general experience of caisson works as well as our own experiments

with animals are distinctly at variance with these results. We have
noted only two instances of what may be called " chance incidence " of
paralysis: (a) goat XV (Series II) had a paraplegia after 15 minutes'
exposure at 75 lbs. and decompression in 30 minutes uniformly, and after-
wards lived for some time, passing through much more severe experiments
without symptoms, and eventually being killed with some difficulty by
75 lbs: 2 hours: 1 min.: (6) goat XIIA (Series III) after 45 lbs.: 1 hour:
10 min. uniform, had apparently very severe bends; it did not however
recover in the usual way and became partially paraplegic, subsequently
passing through many comparatively severe experiments without
symptoms. The tables of our results seem to show quite clearly that as
the conditions of experiment become more searching, not only does the
frequency of symptoms increase but the proportion of severe to total
symptoms becomes much greater.
It is necessary for comparative purposes to form some idea of the
relative importance of these different symptoms, and to consider how
far they may be classed as relatively dangerous or comparatively
negligible. " Bends " are clearly a slight symptom ; there is abundant
evidence both in goats and men that their occurrence is no indication of
urgent danger to life. At the other end of the scale we have death.
Dyspnoea is not far removed in significance from death, and lasting
paralyses are somewhat less serious than dyspnoea. Next in order
come pain and those indeterminate conditions which we have grouped as
" indefinite and various general" : these may be followed by death and
are much more indicative of danger than bends. Temporary paralyses
are not so important and we are inclined to the view that they are not
much more dangerous than bad bends. This classification is based for
the most part on our experience as to the kind of experiment with
which each group of symptoms is commonly associated, and the way in
which the different groups are associated together in the same experi-
ment. The individual variability of the animals introduces many
difficulties, but it is certain that the more severe the conditions of
pressure, exposure and decompression, the more likely it is that the
animals will suffer from symptoms which we have classed as severe.
Immunity to symptoms. There is not the slighest ground, either
theoretical or experimental, for supposing that animals or men, as the
result of repeated exposure to compressed air, acquire any immunity to
the formation of bubbles within their persons. It must be remembered
in this connection that the susceptible individuals become eliminated,
so that those who have been through many decompressions necessarily
show more than average resistance. The matter is not so clear with
regard to the exhibition of symptoms resulting from such bubbling.
We have a certain amount of evidence, too vague to be detailed, that
some goats show slight bends rather more easily in their first few
experiences, and it is not difficult to imagine that they might grow to
neglect altogether those bubbles which evidently cause them no very
great inconvenience at any time. With severe symptoms it is of course
different: no one can suppose that a goat acquires immunity to
extensive pulmonary air emboli or to infarction of the spinal cord.
TABLE V.
Series I. June—July 1906. Pressure 75 lbs. positive ( = 6 atmospheres absolute).
Compression in 39—41 minutes. The details of the stage decompressions are shown in
Table IX.
Decompression minutes 40—50 uniform 40-50 stages
Actual exposure minutes 12 15 30 12 15 30
No. of goat Sex
I M 0
Pa M 0 bends 0 0 0
XIII F bends
XVIII F bends 0 0 0
indefini t e 1
XXI F O + O1 bends
XXII F 0 bends
X F bends bends 0
XVI bends bends 0
XX F bends bends bends
XXIV M 0 bends bends 0
XXVII F 0 0 0
XXIX F 0 bends, 0
dyspnoea
XXX bends bends
XXXH bends
III M 0 0
IV M 0 bends
VI M 0 0
XXVI M bends bends
bends
XXVIII F bad
bends
VIII 0
XIV bends
XIX bends
II M bends
XXV M bad
bends
1
Each of these experiments was repeated upon the same animal: both results are
shown. XVIII was generally uneasy, lay down, nothing definite.
Series II. December 1906 to January 1907. Pressure 75 lbs. positive (6 atmospheres absolute); compression 6 minutes.
The details of the decompressions are shown in Table IX.
Actual exposure ming 1 15 30 60 240"
A
Decompression mins. 10 uni- lOuni- 31 31 31 uni- 31 uni- 31 31 31 uni- 31 31 31 70 i 70 uni- 31 31 uni-
No. of Weight form form stages form form stages stages form stages stages stages stages form stages form
goat Sex kilos stages
M 18-3 bad 0 bends 0 bends bad bends
bends bends
M 16-2 bends si. bends j para-1 temp, bends para- bad
jplegia paral. plegia6! bends,
pain S1
M 16-4 bad bends 0 bad bends bends temp, o
bends' bends' paral. o
XI F 17-1 bends 0 bad bends bad bends died11
bends bends
XII M 19-8 died 7
XIII M 17-8 0 0 0 0 bends bends P
XIV F 18-6 0 bends bend?
XV12 M 16-8 si. para- 0 bends bends bends bad •P
bends 18 plegia4 bends
XVI M 10-7 bends Q18 bends 0 bends temp, bends 0 bends bends temp,
paral. paral.
P
XVII F 24-8 0 bends bends, temp.
dyspnoea paral.
o
XVIII F 26-3 bends 0 bends bends 0 obscure3 bad 0 bends i
bends
XIX F 15 0 0 si. bends 0 bends 0 bends 0 0 bends bends
XX F 15-0 0 0 ? bends 0 0 0 0 temp. bends slight died10
paral. bends
XXI 15 F 29-7 bends obscure 2 bends bad bends,
bends dyspnoea
XXII F 19-0 si. bends 0 ? bends 0 died 9
8
XXIII F 21-4 0 bends died
XXIV'« M 35about 00
3
XXV' M 21-5 bends bends 0 temp, 0 bends bends bends bad
bends paral. bends
XXVI F 17-5 0 bad K
bends f
XXVII F 12-0 bends temp, 0
>•
paral. 0 0
2
Animal also lay down and seemed generally ill: no dyspnoea. Lay down, grunted, seemed quite ill: no definite bends or dyspnoea.
3 Lay down and refused to move: no definite local symptoms. 4
f Could walk with feet dragging in 6 days; nearly well when killed 9 weeks
66 6 7 8 03
later.
fttpr. Well
Well in 9
Q days.
Havs. 6 Almost
Almnst well
wpll in 48 hours.
hnnrs. — . - in
7 Died . - 15
- minutes.
. „ Died
_ . . in
. . 30
- minutes.
. » . Died
. _ . . in
. 76 minutes: bad
10 u ia
bends, dyspnoea, no paralysis. Pain, paralysis, dyspnoea; died 43 minutes. Dyspnoea; died in 26 minutes. 72 lbs., 4 hours,
13
40 seconds; convulsions, dyspnoea, died 20 minutes. 75 lbs., 2 hours, 45 seconds; bends, convulsions, dyspnoea, died 17 minutes.
14 15
Injured shoulder: killed immediately after decompression: see p. 412. 75 lbs., 1 hour, If minutes; died in 10 minutes without
16
dyspnoea or any symptoms except collapse. 80 lbs., 2 hours, 10 minutes uniform; bends, paralysis, dyspnoea; died in 12 minutes.
17
In some of these experiments slow compression in 39 minutes was used, in which case the actual exposure was reduced by half the time
18
of compression. A somewhat different decompression was used in this experiment, viz. 75 lbs. to 17 lbs. in 5 minutes, wait 5 minutes,
then 5 minutes at 13 lbs., 10 minutes at 9 lbs., and 10 minutes at 4 lbs.
TABLE VII.
CO
Series III. February to June 1907. Pressure 45 lbs. positive (4 atmospheres absolute). The details of the stage
decompressions are shown in Table IX.
Exposure m i n u t e s 15 30 45 60 90 240 480
n minutes.. 2 2 30 30 11 10 uni- 30 52 30 l 10 30 10 uni- 30 10 uni-

No. of Weight stages stages form stages stages uniform stages stages form stages form
goat Sex kilos
M 24-3 0 0 0 0 0 bad oends 0 0 0 0 bends 0
3 bends
M 19-6 0 0 0 0 0 0 para- 0 0 0 0 0 0
4 plegia 4
M 17-0 0 0 0 0 bends 0 si. temp. temp, 0 0 bends 0 temp,
XA bends paral. paral. paral.
M 20-0 0 bends 0 0 0 0 0 0 bends 0 0 0 j 0 0
XI A F 19-2 0 0 para- 0 0 0 0 0 0 bends temp,
XII A plegia 2 paral.
F 28-2 0 0 bends bends bad bends
XIII A l a bends dyspnoea
para-
plegia7
XIV A 8 M 16-4 0 0 0 0 bends 0 Si. 0 bends
bends
XV A F 15-5 0 0 0 0 0 0 obscure 3 para- bends bends 0 bends bends bends bends I
plegia 5
XVI A M 21-1 0 bends 0 0 bends 0 0 para- 0 0 0 temp,
plegia6 paral.
XVII A9 F 26-8 0 0 0 0 n 0 0
i1
XVIII A " F 32 0 0 0 0 0 0 0 0 0 0 0 bends bends
XIX A M 22-4 bends 0 0 0 0 0 bends si. bends bends 0 bends 0
bends i
XXA'« F 19-8 0 0 0 0 0 0 0 0 jbends
XXI A F 16-0 0 bends 0 bends 0 bends bends bad bends 0 bends bends bends bends; bends
bends I
XXIII A F 26-3 0 0 0
1
Six other animals had this experience also, of which 7, XXVII A, and XXIX A had bends, and XXIV A, XXX A and 9 showed
no symptoms. 2 Never so completely paralysed that it could not walk, but no improvement occurred during 21 weeks. 3 Bleating,
4
seemed lost and walked into objects as if blind: well in 2 hours. Acute distension of stomach developed next day; goat killed.
6 6
Killed next day. Hind legs and, to a less degree, fore legs also: could walk in 3 days ; killed 15 days when much improved.
7 8 9 10
Killed 3 days. Broke leg, killed. Died of pneumonia: no reason to connect this with decompression. Died, cause
u
unknown. 75 lbs., 3 hours, 58 seconds: died 12 minutes without showing dyspnoea or any symptoms except collapse.
12
Apparently elderly.
TABLE VIII.
Series IV. January to June 1907.
Pressure lbs. positive. 75 51 45 39 30 25 20
Exposure minutes 3 10 30 60 120 180 180 120 120 240 240
Decompression minutes .. l 1 l 60 uni- 90 uni- 68 68uni- 86 92 100 134 134 75to24inl£ t o - 6 lbs. t o - 6 lbs. .0 uni-
No. of Weight form form stages form stages stages uniform stages uniform mins., wait in in form
goat kilos* 1 hr., total 6 mins. 7 6£mins.'
3 19-2 0 sends 0 0 bends 118 or 131 (t) 1 0
4 20-0 0 0
XA 17-8 para- bends temp. 0 0
plegia11 paral.
XI A 23-0 0 0 0 0 0 sl.bends 0
XII A 19-2 0 0 ! 0 0 bends bends
XIII A 28-2 0 temp, paral., 0
paral. dyspnoea
XIV A 16-4 0
XV A 15-5 bends bends 0 0 0 lends 0
XVI A '22-6 0 0 bad died 6 0 0
bends
XVII A 26-8 0 0 P
XVIII A 32-0 0 0 bends bends 0 0
XIX A 25-3 0 bends bends bends 0 ?bends bends -0 O
XX A 19-8 bends 0 0 0 0
XXI A 16-0 0 bends bends bends 0 0 0 0 bends 0
XXIII A 26-3 0 0 0 para- 0 0
plegia12
XXIV A 14-4 0 temp, 0 0 . bends bends temp. 0 ? bends bends, 0 0
paral. paral. dyspnoea
7 19-0 0 0 0 bends bends t bends bends dyspnoea 1 bends
XXV A 14-6 0 died"
XXVI A 14-4 bends 0 0 0 0 0
XXVII A 27-6 0 died 3 0 0 bends 0 0 bad 0 0 bad 0 bends
bends bends 13
XXVIII A 21-1 bends pain 8 , 0 0 bad bends died14 0
temp.
paral.
XXIX A 18-5 0 0 bends pain, 0 0 0 indefi- bends, 0 0 32
dyspnoea nite 10 dyspnoea
XXX A 20-6 0 0 0 +0 0 bends, 0 0
dyspnoea;
8
. 9 26-6 obscure 0 0 0 . 0
t bends 0
bends, 0 bends
dyspnoea
5
> •
XXXIIA '• 22-0 I bends + 0 died 0 0
XXVII 12-0 0 0 0 0 0 bends
1 to
None of the animals in these two experiments showed any symptoms while watched for 1 hour at 24 lbs. and for £ hour at 17 lbs. The results in the
2 CO
Table are those observed after final decompression. Some of these weights differ from those in Series III, Table VII: several of the animals became some-
3
what thinner in the later part of the experiments. Bleated, some convulsions, fore and hind legs paralysed, a little dyspnoea; found dead next morning; 05
4
no bubbles. Cried out, lay down, refused to move, evidently very ill but no local symptoms or dyspnoea; died between 17J and 18 hours later; bubbles
5 6
in heart. Bleating, dyspnoea, died 27 minutes. Paraplegia, dyspnoea, died 190 minutes after decompression and 130 minutes after returning to atmo-
7
spheric pressure. In these experiments the pressure fell to atmospheric pressure in 30—35 seconds, the thermal effect then produced a delay of about
2 minutes when the pressure began to fall further, reaching - 6 lbs. in about 4 minutes more: decompression was therefore in rough stages. The animals were
8
maintained at - 6 lbs. for 1 hour. No symptoms for 18 minutes, then cried out, was moderately convulsed and right fore leg appeared paralysed, no
9
dyspnoea, seemed very ill for about 10 minutes but then recovered and showed no symptoms later. Constant nystagmus also noted lasting about J hour.
10 u 12 13
Lay down, evidently ill, no definite symptoms. No improvement, killed 4 days. No improvement, killed 6 days. On a second trial again
14 14 16
showed severe bends. Castrated, of male habit. Paralysis, nystagmus, no apparent dyspnoea, died 24 minutes. An aged animal.
6. Results of goat experiments.

The detailed results of the experiments on goats are set out in the
accompanying tables. With the exception of those in Series I, these tables
contain nearly all the experiments which were made. Read vertically
the columns give the results of the different combinations of pressure,
exposure and decompression: the records of individual goats can at the
same time be read on the horizontal columns. The tables give however
no indication of the chronological sequence of events. In Series I only
a few experiments are given; the series actually comprised 164 experi-
ments on 34 animals, but the procedures adopted were, through
ignorance, so ill-devised that no very definite results were obtained,
though we gained information which enabled us to devise more satis-
factory experiments subsequently. We have therefore extracted from
Series I only a few results which illustrate the difference between stage
and uniform decompression. The four series roughly represent four
batches of goats, except that the animals of Series IV are the remnant
of Series III with the addition of a further small herd. When reference
is made to individual goats of Series I, the series is noted; otherwise
the goats of Series III and IV are distinguished by "A": this does not
apply to hornless animals which are specified by Arabic instead of Roman
numerals, no two goats having the same number.
TABLE IX. Showing the decompressions of goats from, 75 /6s. and

45 lbs. Times given in minutes, one minute being occupied in each
drop after the first.
Decompressions
Decompressions from 75 lbs. + from 45 lbs. +
X -N
Series II II II IV IV IV IV* ''ill III III
4, 5. 8. 9. 4 3, 4, 7, 9
Columns.. 4, 5, 6 11,12,13,16 (part) 14 12 7 10 9 12,15 8 13 16
First drop in 4 4 5 2 3 3 3 4 1 2 2 2
Wait at 32 3
.. 27 5 2 7 15 4 9 2
,, 22 J 4 5 19 4 14 4
„ 18 4 2 5 5 19 9 14 16 4 9 14
., 14 4 4 4 10 24 9 14 16 3 14 14 14
„ 9 4 or 9 4 9 15 24 14 14 19 9 14 14 14
,, H 10 or 14 10 9 20 24 19 14 19 14 14 14 14
Total time 41or50 31 36 70 134 68 92 86 30 52 57 62
In the next table (Table X) the results are condensed and grouped
in a simpler way, and one or two more experiments are given from
Series I.
i to to - j Pressure lbs.
i O Cr< en positive
OS
Compression
minutes
HtOUHHPHPHH
icWMtctfocifliOi^MHoo^ii-ooactsioiciobaciOJwwwHMHMMH Actual exposure
IS%\ o oo o O O O O O O O O O a i O 0 I O C O O O O O O O O O O O O O 0 » a < U T C n 0 < O ( ? i W C 0 W minutesF
6 6 i
I—' )—» OS CC h-' Ol CO 4 K § Decompression
c B B o
O CD M o
H M O5 CO C
B S 3
«3 CD ?*• r** r1" t3 r1"
2. O" &
2 r 3£ .rfko « to © *>• o c ^ os to ui © oc iSo. of goats P

»
O
I Percent.
B L
IIB.
CD
Cn 05 OD t>3 t o O3
8 g S
B 5' g
>o c•
CD GO
S. P
Temporary 02
§ ? paralysis
Various
indefinite
Paraplegia
Dyspnoea
T
O )-» O © M ( W©H-H-H-©OMO© g°ymptoms
CO
H- H* ©tf*.< O
The following tables give in the simplest form the experimental
evidence on certain poiuts which are of especial importance.
(I) Experiments showing that a certain minimum pressure is required to
give symptoms in goats, and that the results vary with the pressure.
TABLE XL
Pressure in Exposure Decompression No. of No Severe
lbs. positive in minutes in minutes goats symptoms Bends symptoms Death
20 240 2 22 21 1 0 0
25 240 2 23 21 2 0 0
30 60 10 uniform 19 15 4 0 0
45 60 10 „ 11 7 3 1 0
60 1 45 15 4 1 3 0 0
75 15 31 „ 36 19 13 3 1
75 2 50 10 „ 4 0 0 0 4
1
Experiment in Series I : compression SO minutes, exposure 30 minutes.
2
Series I : compression 40 minutes, exposure 30 minutes.
These experiments show that the effects become more severe as the
pressure increases although the duration of exposure was at the same
time diminished and the duration of decompression increased. It was
necessary to arrange the experiments in this way to prevent an incon-
venient mortality among the animals.
(II) Experiments showing that the duration of exposure to high
pressures is of great importance.
TABLE XII.
Pressure 75 lbs. positive, reached in 6 minutes.
Exposure Decompression No. of No Severe
in minutes in minutes goats symptoms Bends symptoms Death
1 1 6 6 0 0 0
3 1 5 4 0 1 0
6 1 6 6 0 0 0
10 1 7 6 0 1 0
15 10 uniform 7 2 3 1 1
15 31 stages 34 29 5 0 0
30 31 „ 23 12 8 3 0
60 31 „ 22 15 4 3 0
120 31 .. 9 0 7 1 1
240 31
These experiments show that goats have taken up enough air in

15 minutes to give severe symptoms on decompression in 10 minutes,
while, if the exposure is less than 10 minutes, nearly all the animals
escape, even with sudden decompression. Note too that with short
A. E. BOYCOTT, Q. C. C. DAMANT AND J. S. HALDANE 397
exposures and rapid decompressions such symptoms as appear are more

frequently severe, and that bends are proportionately less common than
with longer exposures and slower decompressions. Beyond 15 minutes
exposure the results are somewhat irregular, but on the whole there is
a progressive increase of bad symptoms up to two hours exposure. The
results after four hours exposure are about the same, but the animals
used (see Table VI, Series II) were to a large extent selected by previous
experiments, so that it would appear that goats are practically saturated
in about three hours1.
TABLE XIII.
Pressure 45 lbs. positive.
Exposure Decompression No. of No Severe
in minutes in minutes goats symptoms Bends symptoms Death
15 1 15 14 1 0 0
30 1 15 12 8 0 0
60 1 14 10 4 0 0
120 1 10 4 2 4 0
60 10 uniform 11 7 3 1 0
120 10 ,, 11 6 4 1 0
240 10 „ 11 6 4 1 0
460 10 .. 11 6 3 2 0
These figures show that with a duration of exposure up to about

three quarters of an hour, no severe symptoms follow even sudden
decompression. The series with sudden decompression shows that the
results after two hours are much worse than after one hour. This is
not clear from the series with 10 minutes decompression, which, how-
ever, show that the results do not become distinctly worse even after
1
The following figures have been compiled from the records of Heller, Mager and
von Schrotter as illustrating the saturation time for dogs of about 10 (?) kilos. The corre-
sponding data for other animals do not seem to have been determined. Pressure 62— 69 lbs.,
compression in 5—16 minutes, decompression J to 1 minute.
Mild Paralysis
Exposure Number of No paralysis Lasting and
minutes experiments symptoms and bends paralysis asphyxia Asphyxia
Less than 10 1 1 0 0 0 0
10—29 6 5 1 0 0 0
30—59 12 0 6 2 3 1 (lived)
60—120 19 2 2 2 3 10
Four of the group '' paralysis and asphyxia " died, and the other two would probably
have died if they had not been killed. All but one in the " asphyxia " group died, but in
none of the rest was the decompression immediately fatal. These results seem to show
pretty clearly that dogs require more than an hoar to become saturated. It is strange
that the authors conclude (Luftdruckerkrankungen, p. 806) that saturation is so far
complete in about 38 minutes in man that no further intake of nitrogen is of any practical
importance.
eight hours exposure. Note that of the nine severe symptoms, five
were temporary and four permanent paralyses: only one case of
dyspnoea was seen in the whole of the experiments at 45 lbs. (Series III)
and the one case of severe illness of obscure nature was suggestive of
temporary local cerebral anaemia. At 75 lbs., out of 26 severe cases,
four had dyspnoea, four permanent and 12 temporary paralysis, and six
indefinite: seven died.
(Ill) Experiments to show that the duration of decompression
is of great importance.
TABLE XIV.
Pressure 75 lbs. positive, reached in 6 minutes.
No. of No Severe
Exposure Decompression goats symptoms Bends symptoms Death
15 10 uniform 7 2 3 1 1
15 31 „ 36 19 13 3 1
15 90 ,, 12 9 3 0 0
30 31 stages 23 12 8 3 0
30 68 „ 14 14 0 0 0
120 31 ,, 9 • o 7 1 1
120 92 .. 19 15 3 1 0
(IV) Experiments to show that the absolute range of pressure through which
decompression occurs may be of less importance than the relative range
of absolute pressure.
TABLE XV.
ssion
§
c 1 H •sis 1
posure i
11 of prt
i. of goa
L minut<
[ration i
A
3compr<
"8J
inutes
i ft
a
lbs.
•£ = g oo
p 3§i
s u
s? 'C
a
»g
1
£~ H S
p 2
X °" o n n
S*;
75 180 + 24 51 2-3:1 1J 10 10 0» 0 0
51 180 0 51 4-4:1 4 10 2 3 3 2
45 120 -6 51 6-7:1 6 3 0 1 1 1
39 120 -6 45 60:1 6 4 1 0 3 0
45 120 0 45 4-0:1 1 10 4 2 4 0
1
There were three cases of bends at the ultimate end of a two hours' decompression.
A sudden drop of about 50 lbs. from 75 lbs. positive to 27 or 24 lbs.

positive has been made about 200 times altogether in the course of
these experiments without producing any symptoms, and about two-
thirds of the animals showed no symptoms at the end of the stage
decompression. The animals were however only left a short time at
27 lbs. before proceeding with the further decompression. In the
A. E. BOYCOTT, G. C. C. DAM ANT AND J. S. HALDANB 399
present series the animals were left for one hour at 24 lbs. and watched
very carefully, and afterwards suddenly decompressed to 17 lbs. and
again observed for half an hour. The same goats were subsequently
dropped suddenly from + 51 lbs. to atmospheric pressure with very
disastrous results, and a drop of 51 lbs. from +45 lbs. to —6 lbs. was
even worse. The details of these experiments are given in Appendix III.
Owing to the cooling effect of rapid decompression, the falls from +45
and + 39 to — 6 lbs. were interrupted by a delay of about two minutes
at atmospheric pressure so that they were in a rough way stage
decompressions.
(V) Experiments showing the importance of the mode and spacing

of decompression.
The next table shows in brief the results of seven groups of experi-
ments undertaken with the purpose of directly testing the results of
stage decompression in comparison with those of uniform decompression
in the same total time. The only exceptions to the parallelism of the
experimental conditions are (1) in group f the time of uniform decom-
pression was extended from 92 to 100 minutes in order that it might
correspond to the supposed safe rate of 20 minutes an atmosphere1; and
(2) in group /3 stage decompression, three animals were decompressed
by stages in an abnormal way (see Table VI and note, Series II);
since these stages were certainly not more favourable to the animals
than those used for the rest of the group, we have included the results.
In considering these results it must be clearly understood that the
stage decompressions used were not in most cases intended to be safe
for the particular exposure to which they were attached. The only two
groups which were intended to be safe (8 and rj) gave fairly satisfactory
results; with 30 minutes exposure (+ 6 minutes compression) at 75 lbs.
and 68 minutes stage decompression, we obtained no illnesses in 14
goats, and with three hours exposure and two and a quarter hours stage
decompression only two cases of bends in the same number of animals.
For comparative purposes it was desirable that the stage decompressions
should produce symptoms of some kind, and they were intentionally
designed so to do in so far as our knowledge allowed2.
1
The details of the experiments in this group are given in Appendix III.
2
The stage decompressions from 45 lbs. pressure are likewise all shorter than what we
calculate to be safe. The stoppages are also imperfectly spaced. The proper spacing and
duration of stoppages could not be calculated till the results of the experiments were known,
and we realised the extreme slowness of saturation and desaturation.
TABLE XVI. Showing the comparison between the results of stage

and uniform decompression. Pressure 75 lbs. positive.
A. All experiments.
S Bends -
U 00
Jl II I!
S
2
o ft •I
12-30 45 s 18 22 12 55 9 10
/3 15 31 18 34 29 85 2 J>
30 31 15 23 12 52 7 8
30 68 14 14 14 100 0
120 70 14 14 9 64 4 4
120 92 19 19 15 79 3 3
180 134 14 14 12 86 2 2
Total 140 103 74 0 2 27 3 32 4 0 0
12-30 45 uniform 19 32 14 44 15 16
15 31 18 36 19 53 8 13
y 30 31 6 6 1 17 3 4
5 30 68 14 14 7 50 7 7
e 120 70 13 13 4 31 6 7
£ 120 100 19 19 10 53 1 3
•n 180 134 10 10 5 50 1 3 5
Total 130 60 46 3 5 43 4 55 4 3 13
B. Experiments on identical animals.
p 15 31 stages 18 34 29 85 2 2 1 5 0
y 30 31 6 6 4 67 2 2 0
5 30 68 14 14 14 100 0 0
e 120 70 13 13 9 69 4 4 0
t 120 92 19 19 15 79 3 3 1 1
V 180 134 10 10 8 80 2 2 0
Total 96 79 82 0 2 13 1 16 1 0 0 0 1 0
|8 15 31 uniform 18 36 19 53 2 3 8 13 1 2 3 1
y 30 31 6 6 1 17 3 1 4 1 1
5 30 68 „ 14 14 7 50 7 7 0
e 120 70 13 13 4 31 1 6 7 1 1 2
i 120 100 19 19 10 53 1 2 3 2 1 2 5 1
V 180 134 10 10 5 50 1 1 3 5 0
Total 94 46 49 3 5 28 39 2 11
1
Group o compressed in 39 minutes; the rest in 6 minutes or, with long exposures,
in 39 minutes and half the time of compression deducted from the actual time of exposure.
The results are given in two forms: (A) shows the fate of all the
animals tested to obtain the direct comparison between stage and
uniform decompression, while in (B) the figures are confined to the
effects (in the same experiments) on animals which were exposed to
both stage and uniform decompression in each group. This emenda-
tion removes the only very severe symptom and three out of four of the
temporary paralyses caused by stage decompression. Goat XXI
(Series II) was advancing in pregnancy and, after having nearly died as
the result of quick stage decompression, was excluded from the experi-
mental troupe; the effect of the corresponding uniform decompression
on this animal can therefore be only surmised. The effects of
31 minutes stage decompression after 30 minutes exposure were so bad
that, not wishing at this stage to risk losing any animals, the parallel
experiment with uniform decompression was limited to the more
resistant animals. In group /3 (B) two of the animals were only
decompressed once by stages. One had died from uniform decompres-
sion, and the other had broken a leg and had to be killed.
The figures show that the ratio of animals showing no symptoms
with stage decompression to those escaping after uniform decompres-
sion in the same total time is about eight to five. Be it noted too that the
difference between the two methods is in the same sense, i.e. in favour
of stage decompression, in each of the seven groups, including group a
(Series I) where the stages were less well arranged than afterwards. The
difference between the two methods appears still more strikingly in the
quality than in the quantity of the symptoms produced. For while but
one animal had symptoms which can be called distinctly severe after
stage decompression, as many as eleven were materially ill after the
corresponding uniform decompressions, and one died.
This difference may perhaps obtain more definite expression if we
assign numerical values to the different symptoms. Making bends = 1,
temporary paralysis = 2, and so on up to death = 6, we obtain the
following results, showing a ratio of nearly five to one (B grouping) in
favour of stage decompression :—
Group Stages Uniform
5 30
2 6
0 7
4 13
5 26
2 5
Total 18 87
This method is of course very rough. " Death " is worth more than
six times " bends," and bends should have different values according to
the sort of experiment. Bends arising from short exposures and
relatively rapid decompressions {e.g. group /3) indicate that the
exposure has been long enough to allow material saturation aud
are very significant, while if bends show merely the extreme slow-
ness with which the tissues in which they arise get rid of the
excess gas {e.g. group rj), they are of much less moment.
If we exclude bends, and count only the more serious symptoms, or
death, the comparison becomes still more striking, the ratio being then
two for stage decompression, as compared with 50 for uniform decom-
pression.
(VI) Experiments illustrating the difference between different

kinds of animals.
(1) Five goats (XXIV A, XXVI A, XXVII A, 7 and 9), 10 small
guinea-pigs (175 to 275 gms., average 230 gms.), 9 mice (average
20 gms.), 12 small rats (average 35 gms.), 9 medium rats (average1
85 gms.), 8 large rats (average1 200 gms.), and 4 rabbits (1285, 1450,
1850, 2850 gms.) were compressed to 72 lbs. in 7 minutes and left at
that pressure for 3 more minutes and decompressed in 50 seconds.
Goat 9 had a curious short seizure and rolled over on the ground 10
minutes after decompression ; it seemed alright immediately afterwards
and showed no after effects. One small rat became paraplegic at once,
and two other small rats were found dead next morning; one of these
had bubbles in the heart. The rest of the animals showed no
symptoms. The incidence of illness on the young rather than on the
old rats is curious in view of the demonstration of the general im-
munity of young animals by Hill and Greenwood2: it was perhaps
correlated with the shortness of the exposure.
(2) Twelve small rats, 13 medium rats, 8 large rats, 59 mice,
7 rabbits, 10 guinea-pigs, and 1 old hen were raised to 72 lbs. in 10
minutes, left for 1 hour and then decompressed in 50 seconds. No
goats were put in since it was well established that this experience
would have killed all of them. The hen and the largest rabbit (weight
2800 gms.) died in 5 minutes, and 1 guinea-pig became paraplegic
in 10 minutes and died in 20 minutes. All three were extensively
1
The details were eaten by a goat. All the animals were about the same size.
2
British Medical Journal, June 22nd, 1907, Supplement, p. 408.
bubbled; it is interesting to note that there were no bubbles in the

avascular eggs of the hen. None of the other animals showed any
symptoms.
(3) Five rabbits, 10 guinea-pigs, 23 mice, 10 small, 9 medium and
6 large rats were compressed to 51 lbs. in 7 minutes, left there for
2 hours 56 minutes and decompressed in 45 seconds. In similar
experiments, out of 10 goats 2 died and only 2 escaped without
symptoms (see above, p. 398). The largest rabbit, a very fat animal
weighing 29 kilos, died 9 minutes after decompression : the rest
showed no symptoms.
(4) Six goats (3, XIA, XIIA, XXI A, XXIV A, XXVIIA), 7 guinea-
pigs and six rabbits were raised to 75 lbs. in 5^- minutes, left for 15
minutes and decompressed in 42 seconds. Goat XXVIIA had dyspnoea
and paraplegia and was found dead next morning: XXIV A had
temporary paralysis of both hind legs without dyspnoea and was quite
recovered in an hour: 3 and XXIA had bends, while XIA and XIIA
showed no symptoms. None of the small animals were affected.
(5) Seven goats (3, X A, XI A, XII A, XXI A, XXIV A, XXVII A),
7 guinea-pigs, 5 rabbits, 7 medium and 12 large rats and 37 mice were
compressed to 75 lbs. in 6 minutes, left 10 minutes and decompressed
in 48 seconds. Goat X A had paraplegia. The other goats and the
small animals showed no symptoms.
(6) Guinea-pigs, mice and rats were compressed with ourselves to
30 lbs. in 15 minutes, and 1 guinea-pig, 1 mouse, 1 medium and 1 large
rat were killed with chloroform after 33 minutes. After decompression
in 26 minutes by stages, many bubbles were found in the heart and
vessels of the guinea-pig, a few in the mouse's heart, a few in the great
vessels of the large rat, but none in the medium sized rat.
7. Individual variation among the experimental animals in their

susceptibility to decompression symptoms.
The variation in the individual susceptibility of different goats is
very marked. The same variation has been noted constantly among
both divers and caisson workers, and is apparent in most of the
published animal experiments. As an example, the following figures
have been extracted from the tables of experiments at 75 lbs. All four
animals were males and very similar to one another in all obvious
respects: two were resistant and two susceptible.
TABLE XVII.
XIII X
Exposure Decompression (178 kg.) (164kg.) 2 (16-2kg.) XV (168 kg.)
15 mins. 31 mins. uniform 0 0 slight bends paraplegia
15 „ 31 „ 0 0 paraplegia 0
15 „ 31 mins. stages 0 0 0 slight bends
15 „ 31 „ 0 0 bends 0
30 „ 31 „ „ 0 bends pain, temporary bends
paralysis
60 „ 31 „ bends 0 0 0
120 ,, 70 mins. uniform 0 bends paraplegia bends
120 „ 70 mins. stages 0 0 bends bends
In all, therefore, goats X and XIII showed mild symptoms three

times in 16 decompressions, while in the same experiments goats 2 and
XV showed symptoms 11 times, and on 4 occasions these were of a
severe character.
It might be supposed that this variation was only in the exhibition
of symptoms, depending on individual susceptibility to pain, &c, and
did not represent a variation in the amount and distribution of bubbles
within the body. But post-mortem experience shows that the amount
of bubbling present in two animals killed in the same experiment may
be very different; and in living animals it is clear that on the whole
susceptibility to bends involves susceptibility also to the more severe
symptoms, which cannot be much altered by the temperament of the
animal.
The complete explanation of this individual variation in suscepti-
bility probably requires a knowledge of the details of caisson disease
far beyond that which we at present possess. Data exist, however, on
which the influence of several factors may be discussed.
(A) Influence of sex. The following table shows the sum of the
results of the experiments grouped according to the sex of the animals.
The groups denned as " selected" include only those experiments in
which the animals examined were approximately representative: in
Series II for example the figures given are the totals of those experi-
ments in which 10 or more animals were examined, while in Series III
and IV are summed only those observations which included both sexes
about equally (Series III, expts. 1, 2, 5—8, 10—17, Series IV, expts.
18—20).
It would appear from this that there is no clear difference between
the sexes in liability to decompression symptoms in general. The
experiments suggested however that under certain circumstances there
might be a marked difference in the susceptibility to death. In
Series I, of 5 deaths, 3 were in females, a distribution of fatalities

corresponding to the numbers of the sexes (males 12, females 16) used,
while in Series II are shown 1 death in 7 males and 4 deaths in 11
females. All these last four animals were to some degree advanced in
pregnancy, and their mortality is very probably to be associated with
this condition, which, in the goat, is accompanied by a marked increase
in the subcutaneous and intra-abdominal fat. That the deaths in
Series I did not fall more heavily upon the females is perhaps to be
correlated with the fact that these experiments were made in the
summer and none of the goats were found pregnant, while the
autopsies of Series II showed that in the winter practically every
female is pregnant.
TABLE XVIII.
Males Females
Decom- Per Decom- Per
Number pressions Illnesses cent Number pressions Illnesses cent.
Series I :
Total 12 78 25 32 16 71 35 49
Series I I :
Total 7 84 42 50 11 91 38 42
Selected 7 64 26 41 11 79 27 34
15 mins. ) 28 9 32 11 42 13 31
7
exposures J
1 and 2 hrs.)
exposures j 7 25 11 44 8 24 10 42
Series HI and IV:

Selected 7 108 26 24 8 113 29 26
Total 26 270 93 34 35 275 102 37
Influence of size. In the same way the influence of size on sus-

ceptibility may be examined. In the next table the animals are
grouped as above and below the average weight for each sex.
TABLE XIX.
Above average weight Below average weight
Decompressionsi Illnesses Per cent. Decompressions Illnesses Per cent.
Series I I :
Selected 55 22 40 80 26 32-5
15 mins. ex-) 33 39 9 23
27 9
posure J
1 and 2 hrs.) 12 46
9 45 26
exposure }
Series IH and IV:
Selected 102 19 19 119 36 30
Total 157 41 26 199 62 31
The results are contradictory unless (which appears hardly possible)
there is an essential difference between exposures to high (75 lbs.) and
to low (45 lbs.) pressures. The sums of the whole show no material
difference between large and small goats. Theoretically, with decom-
pressions of moderate length such as were used in the experiments
under consideration, small goats should be somewhat more susceptible
than large goats with short exposures since they should saturate more
quickly in proportion to their relatively greater gaseous exchange1.
This is not borne out by the experiments, in which however the rate of
decompression may not have been quick enough to bring out the
difference. It is, on the other hand, obvious that the larger goats
should be more susceptible after long exposures with any except very
short or very long decompressions: this is confirmed by the experi-
ments at 75 lbs. (Series II), but those at 45 lbs. (Series III) show a
greater difference in the opposite sense.
In comparing the incidence according to sex with those arranged
according to weight, it will be noted that in Series II the males are
somewhat more susceptible though they are rather smaller, while the
same experiments, arranged by weights, show that the heavier animals
suffer more frequently. In Series III, in which the male group is
again composed of smaller animals, the susceptibility of the sexes is
equal, while the lighter animals are more susceptible if weight be taken
as the criterion. The only conclusion to be drawn is that these figures
do not indicate that either sex or weight was a determining factor in
the incidence of decompression symptoms.
Influence of the activity of gaseous exchange. General considerations
suggest rather strongly that the susceptibility would be found to vary
with the activity of gaseous exchange, directly as regards short ex-
posures and inversely as regards long exposures. In most of our
experiments, especially those of Series III, the incidence of symptoms
has been conditioned rather by the mode of decompression than by the
duration of exposure. As a whole, then, the goats with the most active
exchange should prove to be the least susceptible.
The respiration results already given have been analysed in reference
to this point: the results are variable and inconclusive and need not be
detailed. This is perhaps not very remarkable when we consider that
the animals were not grouped for the respiration experiments according
1
The respiratory activity per unit of body weight, being proportional to the ratio of
surface to mass, would of course vary but little in the goats, and would only be about a
fourth greater in a goat of 15 kilos than in one of 30 kilos.
A. E. BOYCOTT, O. C. C. DAMANT AND J. S. HALDANE 407
to their susceptibility but by the bands into which they had been
marshalled for the pressure experiments. A factor of considerable im-
portance, which is to a large extent beyond control, is the activity of
the goats at the moment. Some goats are naturally vivacious while
others are almost constantly lethargic. These individual idiosyncrasies
are no doubt of some moment in relation to susceptibility, but the
customary habits of a group of animals may be altogether upset by an
incompatible companionship in the chamber during an experiment.
One group of measurements gave for example the results shown in
the following table. The CO2 production of each group was deter-
mined on four separate occasions under conditions similar to those
obtaining in the pressure experiments.
TABLE XX.
Females. Males
CO2gms.
No. of Weight 111- Pressure per kilo No. of Weight Ill- Pressure per kilo
kilos Expts. nesses lbs. per hour goat kilos. Expts. nesses lbs. per hour
XIIA 14-2 14 3 19-2 15 3
0 0-921 1
XV A 15-5 17 4 20-0 16 0 0-669
45 1-104
XVIIIA 31-4 17 2 XA 17-8 17 6 45 0-697
45 1-020
XXIA 170 17 11 XIA 23-0 17 2 25 0-959
XXIIIA 22-0 13 0
45 0-852
XVIA 22-6 16 4f 45 0-967
XIX A 25-3 17 7'
Average 20-0 78 25 — 0-974 21-3 98 23 0-823
(32 °/ (23 Inf
The average size in each group is about the same, and the sex
incidence for all goats of Series III is the same. The results therefore
appear to show that the males are 32 °/o less susceptible and produce
17 °/o less CO2—a result which cannot be correlated with theory.
The only experiments made with the animals grouped according to
their susceptibilities gave much more rational results. Great care was
taken in this series to make the conditions as nearly identical as
possible in all four observations; the animals were kept in the dark
and remained quite quiet throughout. The results show a GO2 pro-
duction by the susceptible animals one-sixth less than that of the
non-susceptible.
Influence of blood volume. • The volume of the blood was determined
in 8 goats, in 7 of which the susceptibility to caisson symptoms had
been ascertained. The method used was the simple one of Welcker, in
which, after taking a standard sample of arterial blood, the animal is
bled to death and then thoroughly washed out with salt solution. The
26—2
TABLE XXI.
Non-susceptible goats.
Animals Respiration
CO2 gms. per hour
Total Selected Duration
No. of No. of Bar. in per 1000
goat Sex Weight Decomp.1 111 Decomp. 111 expt. Temp. m. m. hours kilo sq. cms.
4 M 18-6 15 0 10 0
XIA M 20-8 16 1 11 1 I 12° 776 0-717 1-853
XVIIIA P 30-7 14 0 10 0 III 14° 775 0-751 1-941
XXIHA F 25-2 12 0 8 0
Average 23-8 57 1 39 (24°/o) 0-734 1-897
7o)
Susceptible goats.
XVA F 15-0 15 6 10 4
XXIA F 15-4 15 8 10 6 II 14° 775 0-624 1-501
XIX A M 21-3 15 7 10 5 IV 13° 766 0-704 1-693
XIIIA F 25-2 16 5 11 2
Average 19-2 61 26 41 17 0-664 1-597
(43«/0) (41»/0)
1
These figures are given up to the date at which the respiration experiments were
made. Some time elapsed before the final susceptibilities were ascertained: these were
8°/ 0 for the non-susceptible group and 45 °/o * o r the susceptible animals.
tissues were not afterwards extracted with water: the red colouring
matter so obtained is so small in amount that it can have little influence
on the final result, and Douglas1 has shown that additional difficulties
are thereby introduced. For purposes of calculation the specific
gravity of the blood has been taken as 1050, The results and the
decompression records of the goats are given in the two following
tables. The figures should be read in relation to the "clean" weight,
i.e. the crude weight less the weight of the contents of the alimentary
canal, which in these animals is very considerable.
The results seem to indicate that there are two types of blood
volume in goats: one about 1\ °/0 of the clean body weight and the
other about 6J %> the first type being also associated with a higher
percentage of haemoglobin. No relation between blood volume and
susceptibility is apparent; thus goats 2 and XIII, both males, have
identical blood volumes and differ about as widely in their sus-
ceptibility as any two goats which have come under our notice.
Conclusions. Of the four factors considered in detail, it appears
therefore that age, sex and blood volume were without appreciable
influence. Pregnancy and a low rate of respiratory exchange seem to
favour the occurrence of symptoms.
1
Journal of Physiology, vol. XXXIII. (1906), p. 499.
A. E. BOYCOTT, G. 0. C. DAMANT AND J. S. HALDANB 409
TABLE XXII.
Weight of contents of
stomach and intestines Mass of blood
per cent, of Haemo-
Whole Per cent. globin p.C.
No. weight Total of whole Volume of Whole Clean of human
of goat Sex kilos kg. weight blood c.c. weight weight standard>
1 M 19-9 3-8 19-1 1006 5-31 6-56 74
xni M 18-9 4-3 22-7 883 4-91 6-36 —
2 M 18-0 3-7 20-5 874 5-10 6-42 64
X M 17-0 3-0 17-6 833 5-14 6-25 72
XVI M 11-3 2-3 20-3 592 5-50 6-91 64
A F 31-2 5-2 16-7 1874 6-31 7-57 78
XVIII F 27-1 4-6 17-0 1395 5-41 6-47 75
XVII F 24-4 40 16-4 1520 6-54 7-83 84
Average 21-0 3-9 18-8 1122 5-53 6-80 73
Average of males 17-0 3-4 20-0 838 5-19 6-50 68-5
Average iDf females 276 4-6 16-7 1596 609 7-29 79
1
The red blood corpuscles of goats are very small, about 4/J. in diameter (Jolly, C. R.
Soc. de Biol., vol. LXIII (1907), p. 210).
TABLE XXIII.
Pressure 75 lbs.
Exposure Decomp.
minutes minutes Goat 1 2 X XIII XVI XVII XVIII
15 30 un. 0 slight 0 0 bends 0 slight
bends bends
15 30 un. 0 para- 0 0 0 0 0
plegia
15 30 st. 0 0 0 0 0 0 0
15 30 St. bad bends 0 0 0 0 0
bends
30 30 st. bends temp. bends 0 bends bends 0
paral.
60 30 st. bends 0 0 bends 0 dyspnoea 0
120 70 un. bends para- 0 0 bends temp. 0
plegia paral.
120 70 st. 0 bends bends 0 0 0 0
15 1 — bends
15 10 un. bad 0 bends
bends
30 30 un. — — — — bends — bends
30 30 st. 0 0 0 temp. bends
paral.
60 30 st. 0 — 0 0 0 — obscure
120 30 st. bad bad bends bends bad
bends bends bends
240 30 st. bad bends 0 temp. — bends
bends paral.
240 30 un. — temp. — — — .—
paral.
But there are doubtless other particulars which, alone or in combi-
nation, are of fundamental importance. Such other factors we have not
yet been able to examine in detail. Fatness for example can be gauged
only in the dead, and, though we have the distinct impression that the
goats which die easily (i.e. under circumstances of pressure, exposure
and decompression which cause very few severe symptoms and deaths)
are fatter than those which are killed with difficulty, we have had no
means of extending our observations on this head to the great majority
of our animals1. Fatness also involves a low rate of respiratory exchange
per unit body weight. There are grounds in human experience for
holding that age may have an important share in the production of
symptoms and Hill and Greenwood have recently shown2 clearly that
young animals (rats, rabbits and cats) are far less susceptible than adults
of the same species. All the goats used by us appeared to be adult; in
the two cases (XIII A, XXXII A) in which old age had obviously set in,
the susceptibility seemed to be somewhat above the average, but the
ages of the animals as a whole were unknown. In any case such
a factor as old age must be reduced to simpler components before it can
be correlated with the theory of decompression.
8. The pathology of caisson disease in goats.

We have hitherto dealt exclusively with the symptoms exhibited by
the experimental animals rather than with the actual or possible
presence of bubbles within them. We have however made a number of
observations on the post-mortem appearances of goats after decom-
pression, which may be shortly dealt with here. Most of the animals
had died from caisson disease but in other instances they were killed at
varying periods after decompression.
The presence of bubbles in vivo must be inferred from their
discovery post-mortem with considerable caution. The supersaturation
of the body may be such that the separation of the gas as bubbles may
take place after death. There are reasons for supposing that the living
body presents nothing in the way of points or surfaces on which
bubbles might arise in the blood and tissues as they do upon the glass
and dust in soda-water, and a remote analogy may perhaps be drawn
1
We have since examined this point by direct analysis of rats and guinea-pigs divided
into susceptible and non-susceptible groups by decompression experiments. The results,
which will be published in detail later, show that fatness is a very important factor in
individual susceptibility to death.
2
Meeting of Physiological Society, Nov. 1906.
A. E. BOYCOTT, G. G. 0. DAMANT AND J. S. HALDANE 411
with the relations of the vessels to the separation of fibrin intra vitam.
Death may well alter this condition in some degree, and in any case the
time factor is of importance as well as the foreign manipulation which
examination involves. During life no portion of the blood remains in a
supersaturated state for more than the time required for it to return
from the tissues to the lungs. We have already seen that the duration
of this period is probably of very great importance as regards risk of
bubble formation. On cessation of the circulation the blood remains in
a supersaturated state for an indefinite period. In one instance at least
we have actually observed such post-mortem separation of bubbles : a
rabbit was killed immediately after 75 lbs., 2 hours, 31 minutes stages,
opened up at once and no bubbles found1; an hour later a few bubbles
were found in the inferior vena cava. In another case bubbles in the
bladder were seen to increase considerably in number and volume during
the progress of the examination.
The possibility of air being introduced from without into the veins
must also be considered. Ewald and Kobert showed that air might be
forced into the pulmonary capillaries by an increase of intra-pulmonary
pressure such as may occur in severe dyspnoea. We have seen bubbles
in large quantities in the meningeal veins, and in small numbers in the
superficial veins of a fore-foot, under circumstances which left no
reasonable doubt that they had been sucked into the vessels during the
somewhat violent manipulations used in opening the skull and skinning
the leg respectively.
There is not much doubt that some of our animals which showed no
symptoms must have had bubbles present in the blood. Catsaras
decompressed a dog in 1 minute after 2 hours exposure at 65 lbs.: it
showed no symptoms and was killed 6 hours later, when fine bubbles
were found in the blood. Heller, Mager and von Schrotter (pp. 7.90,
882) record two dogs which were killed 10 minutes after sudden
decompression after 65 and 52 minutes at 15 and 18 lbs. respectively: in
both cases bubbles were found in the heart", though there is abundant
evidence*to show that dogs never show any symptoms after decom-
pression from such low pressures. In our own animals attempts were
made to see bubbles in the retinal vessels during life. Though an
1
The vessels and bladder in the rabbit are so thin-walled that bubbles can be seen
with certainty if they are present.
2
On the other hand dogs (showing no symptoms) killed after sudden decompression
after 16 minutes exposure at 2-8 atmospheres, 5+ at 3 "5, 12 at 4 "5 and 5 at 4-7, showed
no bubbles in the blood: these were however found in three other dogs after 10 + minutes
at 4 0 atmospheres, 16 and 72 at 4-5.
excellent view of the fundus may easily be obtained, no bubbles were
ever seen even in animals with severe dyspnoea, so that the method
cannot be taken as giving any indication of the absence of bubbles in
the blood. Some of these animals died and plenty of free gas was found
in the retinal vessels post-mortem. Four animals which showed no
symptoms were killed within ten minutes after decompression with the
following results:
TABLE XXIV.
Results of similar
experiments in other goats
Series and Exposure Decom- Bubbles
number Pressure (actual) pression in Severe
of goat lbs. minutes minutes blood Number Bends symptoms Death
3:XXVI A 25 120 i absent 23 2 0 0
1:11 45 26 6 absent 15 3 0 0
2 : XIV 75 19 31 stages present 29 5 0 0
1:1 78 30 9 absent 4 0 0 1
These goats were killed before the expiration of the appropriate

period for the development of bends. The experience of similar
experiments indicates that they might have shown symptoms if they
had not been killed. Yet three out of four had no bubbles in the blood.
A few observations on rabbits on the other hand gave rather different
indications. Seven rabbits in seven different experiments were exposed
to 75 lbs. for periods of from 15 to 120 minutes and killed immediately
after decompression in 31 minutes by stages. There is no question but
that it would be the very rarest occurrence for a rabbit to have any
symptoms under these circumstances, but in four of the animals we found
bubbles in the heart or great veins. These may however have been
formed post-mortem, and in one such case they were observed to appear
some time after death.
The post-mortem appearances observed may now be shortly described.
It should be remembered that most of the animals dealt with here died
under circumstances of experiment less severe than those of other
observers. The pressure was in almost all cases 75 lbs. and in a majority
of instances the decompression was not instantaneous. These considera-
tions probably afford the explanation of the somewhat less iemphatic
changes which we have noted. The naked eye appearances were in nearly
every case supplemented by microscopical examination of many of the
organs.
Lungs. The amount of blood in the lungs depends upon the
condition of the heart: in severe cases, with the pulmonary artery
choked with bubbles and the right heart distended, they are pale and
bloodless; in other instances the quantity of blood appears about normal.
Haemorrhages may occur and small blood clots are not infrequently
found in the trachea. A very marked, scattered, lobular emphysema is
almost constant; the only explanation appears to be that some bronchioles
are more or less impervious during decompression. Examination of
fresh material showed that bubbles may burst out of the capillaries into
the interstitial tissue of the lungs, and presumably therefore also into
the alveoli. The same was found in rabbits which were killed by the
injection of small quantities of air into the veins. Nothing resembling
the exudative process seen in oxygen poisoning was ever found in animals
exposed to simple air pressures.
In all fatal cases (with one exception) more or less abundant bubbles
were found in the blood. In severe, rapidly fatal cases, the right heart
is much distended with bubbles, the pulmonary vessels plugged with
froth and the left heart nearly empty. The block in the pul-
monary artery may indeed be so complete that the left auricle
is collapsed and puckered up. In other animals, which have lived for
20, 30 or 60 minutes or longer, the two sides of the heart are equally full
of blood and the right heart is not distended. The immediate cause
of death, in all but three cases, which died many hours after decom-
pression, was clearly pulmonary air embolism, and this is doubtless
the cause of the urgent dyspnoea already noted. In two cases death
ensued without any dyspnoea. Both these animals were intention-
ally killed by very severe experiments, viz. 75 lbs. for 1 hour (goat XXI)
and 3 hours (XVIII A) with decompression in 1 | and 1 minutes. Both
showed no symptoms for 5 minutes and collapsed and died quietly in 10
and 12 minutes respectively. Such animals must be regarded as being
so overwhelmed by sudden asphyxiation that they exhibit only the
symptoms of deficiency of oxygen and not those of the accumulation of
carbon dioxide.
The fatal case in which no bubbles were found in the blood post-
mortem was in goat XXVII A. After 75 lbs., 15 minutes, 42 seconds it
showed pain, ate part of a note-book, and became paraplegic. It was
found dead next morning1. In several other cases of delayed death, and
in one (XVI A) in which the animal died in 3 hours2, the quantity of
bubbles found seemed to be altogether inadequate to produce a fatal
result. One may suppose that they had been previously more
1
Heller, Mager and v. Schrotter (p. 852) record a case in a dog fatal in 6 hours after
decompression in which no free gas was found in the vessels.
2
This animal had however been recompressed and died under a state of recompression:
see protocol c, Appendix III.
numerous and that oxygen starvation resulted in death at a time when
the aeration of the blood had been restored. This form of delayed death
from deficiency of oxygen is well-known in e.g. carbon monoxide poisoning.
It might be for instance that temporary obstruction of the coronary
arteries or portal capillaries caused fatal degenerative changes in the
heart muscle or liver cells.
The distribution of the bubbles in the different parts of the vascular
system shows several peculiarities. If only a few are present, all may
be collected in the smaller branches of the pulmonary artery with none
in the right heart. The left heart generally contains a few bubbles; the
amount there and in the arteries roughly varies inversely with the
rapidity of death unless decompression has been very quick. Smallness
in the amount of bubbling affords the heart the best chance of being
able to pass on the froth to the arteries, and cases which die slowly seem
to show distinctly more arterial bubbles than those which expire almost
at'once. The veins contain variable quantities of bubbles, but always
more than the arteries. They are especially abundant in the mammary,
mesenteric, spermatic and portal veins, coronary vessels, and notably few
in the veins on the surface of the stomach and in those of the brain and
spinal cord. In several instances we have noticed great accumulations
of froth in the liver while the spleen at the same time showed no bubbles
in the blood flowing out on section. This massive portal embolism is
probably the cause of the multiple small capillary haemorrhages which
are frequently seen in the omentum and mesentery. Blocking of the
portal circulation might also give rise to general symptoms of a very
serious character.
It should be noted that the liver is particularly badly situated for
getting rid of excess gas during and after decompression ; nearly all the
blood reaching it is already partly saturated by passing through the
intestines, &c. The liver also contains much fat.
Lymph. The lymph in the thoracic duct has been noted to be full
of froth on several occasions.
Other liquid areas. Bubbles have very seldom been found in the
aqueous humour; the blood supply is considerable, so that their absence
is probably to be attributed to the excess gas being carried off during
decompression and the period which the supersaturation phenomenon
adds, for practical purposes, to the actual time occupied in reducing the
pressure to normal. The vitreous humour, on the other hand, has a poor
blood supply and its consistence is such that any bubbles forming there
would remain in situ. On only one occasion have bubbles been found
(goat XXVI, 75 lbs., 2 hours, 31 minutes stages), when they were seen
in a layer close against the ciliary body. Their absence is explicable on
the ground that the vitreous humour would take a very long time to
saturate. The bile often contains bubbles : they were noted in one goat
exposed at 75 lbs. for 15 minutes and killed immediately after decom-
pression in 31 minutes stages, and in 8 animals exposed to the same
pressure for 1—4 hours, but not in two animals exposed for 15 minutes
which died 30 minutes after decompression in 10 minutes and
30 minutes uniformly respectively. The urine found in the bladder
post-mortem is remarkably free from bubbles; on two occasions only has
free gas been found. We have evidence here that the phenomenon
must be due to supersaturation and the absence of "points," since we
have very frequently observed goats pass urine after decompression
which frothed freely on coming into contact with foreign surfaces. This
is often seen in animals which show no symptoms. Thus in one experi-
ment, seven goats were exposed at 45 lbs. for one hour and decom-
pressed in 30 seconds. One had bends 19 minutes later. Within 24
minutes after decompression four animals passed urine; in two cases this
frothed up freely as it ran over the pavement while in the other two
(including the goat which had bends) no bubbles could be observed. It
is somewhat striking to observe the transparent bladder of the rabbit
containing urine quite free from bubbles while the vesical veins coursing
over its surface are full of froth. The cerebro-spinalfluidrarely shows any
bubbles: they have been seen only three times, in all cases after long
exposures (1 to 3 hours) at 75 and 51 lbs. with sudden decompression.
Synovial fluid is almost always full of bubbles; exposure for 15 minutes
at 75 lbs. is sufficient to cause their presence, while decompression in 100
minutes uniformly is not enough to prevent their formation. In animals
which have died within 3 hours of decompression, we have found them
in ever)' case. Amniotic fluid is dealt with below. Bubbles have been
seen after very severe experiments in the pericardial and peritoneal
fluids when present, and in the serous contents of a mammary cyst, but
not in the milk. We have not seen any accumulations of gas in the
serous cavities.
Solid organs. Fat commonly shows bubbles, often in extreme
abundance. They are more numerous in the abdominal than in the
subcutaneous fat; the latter is much more vascular. Other solid organs
for the most part show no bubbles outside the blood vessels-; a few are
sometimes found in the liver and the spinal cord may contain large
numbers. In the liver it is very difficult to determine whether any
bubble is inside a capillary or not, and we have failed to find clear
evidence of bubbles outside blood-vessels, still less actually within tissue
cells, in cardiac or skeletal muscle, spleen, kidney, suprarenal, salivary
glands, thymus, thyroid and parathyroid, pancreas, lymphatic glands,
haemolymph glands, nerves, posterior root ganglia, testis, ovary or
mammary gland, though an extensive systematic histological examina-
tion has been made of more than 20 goats exposed at 75 or 45 lbs. for
from 10 minutes to 4 hours and decompressed in from 30 seconds to
100 minutes.
This condition has no very obvious explanation. It is curious, for
instance, to see the spermatic vein (and sometimes artery as well) full of
froth and yet find no evidence of bubbles in the tissues which it drains:
the same thing is also shown most strikingly in the mammary gland and
vessels. There can be no doubt that these tissues must be fairly com-
pletely saturated in 4 hours and it is impossible that the excess should
be removed from the tissues more quickly than from the blood. It
follows that the blood must stand in an unfortunate relation towards
bubbling in that it effervesces with a smaller difference of pressure,
within and without, or with the same difference of pressure in a shorter
time, than do the more solid organs. It seems unlikely that this
difference depends on the motion of the blood. Rhythmical pulsating
circulation through a smooth elastic system can hardly function as
a shaking which would be efficient in bringing out free gas. Even
if it did, it is not easy to see why the tissues are not affected by the
pulsations in the same way, though perhaps not to the same degree,
since isolated collections of fluid, as in the joints, may bubble very easily.
One can only suppose that the dissolved particles of gas find in the
tissues obstacles, visible and invisible, more obstructive to their aggrega-
tion into bubbles than those occurring in the blood.
Bubbles once formed in the blood will also increase in size more
readily, since their movement will continuously keep them in contact
with fresh portions of supersaturated liquid.
Among the solid organs, bubbles outside the vessels are found most
frequently in the central nervous system. The fatty nature of this
tissue is probably important in this respect. The brain is singularly
free, both by direct examination and by the study of secondary
degenerations. The cord may however contain numerous bubbles, and
a study of their occurrence and distribution gives interesting results.
In the first place they may occur in areas of softening after compara-
tively mild experiments (e.g. 45 lbs., 2 hours, 10 minutes): in this case
2nd cervical. 3rd dorsal.

Fig. 7. Shows the distribution of extravascular bubbles infiveregions of the spinal cord
of goat 3 (series IV). The animal died of oxygen poisoning soon after the beginning
of a decompression of 133 minutes duration by stages after 3 hours exposure at
81 lbs. in an atmosphere containing 36 °/0 oxygen. The bubbles are practically
confined to the white matter and are there especially concentrated in the boundary
zone where the circulation is least good. Each diagram is a composite drawing
showing all the bubbles in 0'4 mm. length of cord.
they are confined to the area in which the circulation has been brought
to a sudden standstill by a collection of gas. On the other hand they
may be found in the cords of animals which have died immediately after
a drastic decompression. This is however rather exceptional. Thus the
cords of three animals decompressed in less than a minute after 1 hour
at 100 lbs. and 1 and 2 hours at 75 lbs. contained numerous bubbles,
while in two animals treated in the same way after exposure at 7o lbs.
for 1 and 3 hours respectively, none were found.
The distribution of the bubbles when numerous is in harmony with
the theoretical conclusions derived from the blood supply. They are for
instance least numerous in those segments with an abundant blood
supply (lumbar enlargement), and are almost confined to the white
matter, those found in the grey substance being distributed along
its periphery towards the boundary zone between the superficial and
deep vessels. Thus one cord (goat XXI: 75 lbs., 1 hour, If minutes)
contained the following bubbles in 412 cubic millimetres in different
parts:
Grey Posterior Ant.-lat. White
Segment matter columns columns matter Total
2nd cervical 14 141 215 197 175
5th dorsal 11 23 95 87 79
1st lumbar 0 32 161 140 140
4th lumbar 2 15 37 28 19
Average 7 53 127 113 103
Fig. 7 shows the distribution of bubbles in another case: note the

paucity of bubbles in the grey matter and their concentration in the
boundary zone.
The distribution of the areas of softening is also important. With
one exception, these are most marked in, and usually confined to, the
lower dorsal and upper lumbar segments where the blood supply of the
cord may on many grounds be surmised to be at its minimum. They
affect only the white matter. Now the only parts of the body in which
we have found appearances resembling embolic infarction are the white
matter of the spinal cord and the fat. The latter has on several
occasions been found to contain large and small areas of necrosis. We
have obtained no evidence of infarction of the spleen, kidney, heart-
muscle, &c. The distribution of small bubbles by the arterial blood
stream must be universal. They probably lodge in many places: while
they are rapidly pushed forward in the grey matter and in most other
tissues, if they lodge among the fatty surroundings of the capillaries of
the white matter, or in actual fat, they quickly increase in size to such
an extent that their removal becomes impossible. It is also clear that

in consequence of the slow circulation in the white matter, and especially
in such inactive parts as the lower dorsal segments, bubbles have plenty
of time to increase in size in the circulating blood. The condition of
supersaturation will also last much longer in the white than in the grey
matter. The cause then of these areas of softening in the cord is not
ordinary embolism, but embolism which becomes effective to produce
infarction by reason of the effect on the size of the embolus of the local
conditions of the circulation rather than from any of those peculiarities
in the resistance of the different tissues to lack of oxygen, or in the
freedom of collateral circulation, which determine the topography of
common infarcts.
The presence of hubbies in the •uierine contents. We may group
together here a number of casual observations which have been made
on the distribution of bubbles in the foetus and amniotic fluid of
pregnant goats dying of caisson disease. The pressure was in all cases
75 lbs.
TABLE XXV.
Time of Bubbhis present in the
Develop-
Number Exposure Decompression decompression maternal amniotic ment
of goat minutes minutes minutes blood foetus fluid of foetus
XXIII 15 31 stages 30 + 0 0 advanced
XXII 120 31 „ 75 + 0 +
XX 240 31 „ 40 + live + 6 inch
dead 0 4 inch
XXI 60 12 10 + + advanced
XI 240 31 uniform 25 + 0 0 1 inch
XIV 15 31 stages killed at once + 0 0 J inch
XVIIIA 180 1 12 + + + advanced
XXVIIIA 180 4J 24 + 0 0 J inch
XXXIIA 180 4 27 + 0 0 4 inch
(dead) <
These observations seem to be fairly concordant. In 15 minutes the

uterine contents have not taken up much excess of gas (XXIII), nor
does a dead foetus absorb any (XX, XXXII A). In one hour both
foetus and amniotic fluid have taken up abundant excess (XXI), which
may, if death be long delayed after a rather slow decompression, be
discharged from the foetus more quickly than from the amniotic fluid
(XXII)1. With a very young foetus, the circulation is probably too
1
Two pregnant guinea-pigs were exposed for 1 hour at +100 lbs. and decompressed in
34 minutes by stages: they showed no symptoms. On being killed 5 hours later, numerous
bubbles were found in the amniotic fluid but nowhere else.
active and the bulk of fluid too small to favour bubbling (XI, XIV,
XXVIII A).
The amniotic fluid, which contained in this case only a faint trace of
proteid, may show the phenomenon of supersaturation to an exquisite
degree. In goat XXI it was especially noted that a large bubble was
present in the amniotic fluid on removing the uterus from the body.
After free shaking to bring out any more gas, the uterus was opened
and the contents poured into a glass vessel. Contact with this foreign
surface immediately produced a great froth of fine bubbles.
The free gas runs together into one large bubble. Advantage was
taken of this convenient circumstance in two instances to make
analyses. The samples were collected over water and in XXII analysed
at once; in XX they were kept for 20 hours over water before examina-
tion and in this case therefore the figures for OO2 represent minimal and
those for oxygen maximal values.
XXII XX XX
Live foetus Dead foetus
Total gas c.c. 16 27 105
CO2 per cent. 16-23 3-55 2-73
1-10 2 14 0-85
o3 81-90 94-14 96-21
N2 „
Combustible ga.8 (caleu-1 0-17 0 21
lated as CH 4 and H2) } 0-77
These results correspond with those of Bert (pp. 955, 961) of the
free gas in the heart: they are not in accord with those of Heller, Mager
and von Schrotter (p. 800) who found 15"31 and 7-18 per cent, of oxygen
in the free gas collected from the hearts of dogs. It is somewhat
significant that if this excess of oxygen is calculated as an air leak, the
figures of Schrotter correspond exactly with those of Bert and our
<analysis XXII.
Duration of bubbles. It is difficult to say how long bubbles may
remain in the vessels and tissues after their first formation in animals
which survive1. The question is much complicated by the fact that we
have reason to believe that bubbles may continue to form for a long,
and quite unknown, time after decompression. This is probably
especially marked in cases in which either local blocking of the blood
supply has occurred, or the circulation has been slowed generally by a
greater or less degree of cardiac and pulmonary obstruction. It would
1
Zografidi (Revue de Medecine, 1907, p. 159) records the rinding of numerous bubbles
in the peripheral vessels, but not in the heart, of a diver who was paralysed and died
33 days after decompression !
appear likely that bubbles once lodged in the lungs would probably
stop there for a considerable time, since their gaseous composition would
quickly approximate to that of the alveolar air and there would be no
considerable difference of tension to encourage their removal. Such
results as we have which bear on the matter are collected in the next
table. It will be seen that bubbles have been found in the blood of
one animal which died two days after decompression (and that in an
animal which had shown no dyspnoea) and in the joints up to 26 hours.
In the substance of the spinal cord bubbles may persist far longer: in
two cases we have found them 15 days after the last exposure to pressure
and in one 27 days after the last occurrence of symptoms.
TABLE XXVI.
Bubbles present in
r re 88 u re UiXposure Decompression Result
Goat lbs. minutes minutes hours Blood Joints
XXII (Series I) 75 35 40 stages died 39 + _
XXV A 75 120 100 uniform „ 164 + 0
XXVIIA 75 15 „ 15* 0 0
XV A 45 120 4 killed 24 0 0
4 45 120 4 » 26 0 +
XIIIA 45 120 10 uniform „ 72 0 0
XA 75 10 t „ 96 0 0
XXIIIA 45 to - 6 180 6 .. 144 0 0
* Found dead next morning.
A histological point of some practical importance arises in connection

with the size of the bubbles found in the blood. The bubbles soon run
together into large bullae after death so that it is necessary to make
the examinations immediately after death in order to observe approxi-
mately the true state of affairs. It will then be found that there are no
bubbles so small as to be of strictly microscopic dimensions. Nor are
any very small bubbles found in the spinal cord; in any one case all the
bubbles are about the same diameter, commonly some 25 microns. The
same is true of the bubbles given off on decompressing water, salt
solutions, serum, blood, and even such thick solutions as gelatine or
agar. At the same time however it is possible to produce bubbles
which are truly microscopic and which last some hours in some sticky
solutions such as gum and treacle. The energy required to aggregate
particles of dissolved gas into a bubble is evidently considerable, and
there is the same difficulty in the formation of free gas bubbles from
solution in liquids as there is in the separation of liquid particles from
solution in gases and of solid particles from solution in liquids.
Extremely minute bubbles are unstable in the same way as extremely
minute droplets of water condensing from supersaturated air, or salt
crystals forming in a supersaturated solution in water: in all cases the
tendency is to reduce the free energy (surface tension) by reducing the
ratio of surface to mass, and accordingly the smallest bubbles, droplets
or crystals as the case may be, are rapidly, in the case of bubbles
practically instantaneously, abolished to produce macroscopic masses.
This is well seen on watching under the microscope a stream of bubbles
coming off some " point" in soda water. It follows that if the concen-
tration of dissolved molecules of gas is not higher than some unknown
point, bubbles will not be formed. It is possible that the absence of
bubbles from most of the solid tissues is to be explained by this non-
existence of very small bubbles and the mechanical difficulties of the
rapid aggregation of a sufficient number of molecules to produce large
bubbles. It is also doubtless connected with the period of delay in
bubble formation whereby an animal, for practical purposes, gains
several minutes over the actual time of decompression.
It is reasonable to suppose that the temporary paralyses are due to
temporary ischaemia from air bubbles in the vessels. The more lasting
palsies are undoubtedly caused by obstruction sufficiently complete to
produce softening and necrosis. As already mentioned, the change is
confined to the white matter and in nearly all instances affects only the
lower dorsal or upper lumbar region. In these segments the bulk
of cord destroyed may be very extensive : thus in goat XXIII A fully
three-quarters of the lateral columns were destroyed from the eighth
dorsal to the second lumbar segment over a length of rather more than
five inches, and in goat XV A the softening involved nearly the whole
of the lateral columns and parts of the anterior and posterior columns for
a length of four and a half inches. Such cases may recover to a remark-
able degree, and eventually show objective signs of paralysis so slight
as to be hardly perceptible except to those familiar with the individual
animals.
The only other tissue in which we have found any signs of the
results of circulatory obstructions is the intra-abdominal fat. Large
masses of necrosed fat have been occasionally met with, especially in
the fat lying below the kidneys. Only late stages have been seen: the
necrosed areas are then surrounded by a well-marked ring of giant-cell
reaction, and the surface layers are mostly converted into a calcium
soap.
No evidence of infarction in other organs has been seen: the rarity
of such changes seems to afford pretty good evidence that the duration
of any obstruction in such organs as the spleen or kidney cannot be
very long.
The pathological changes underlying the chief symptoms have been
already sufficiently noticed except as regards bends. The exact cause
of bends is not known. They have been attributed to bubbling in the
central nervous system, chiefly on the ground that human expeiience
shows that they are very frequently bilaterally symmetrical. This fact
however cannot be taken as indicating any such origin in view of the
complete symmetry of the limbs (where the symptoms occur) and the
uniform symmetry of the causative agent throughout the body. In two
animals which were killed soon after decompression when they showed
bends only, we could find nothing abnormal in the cord, posterior root
ganglia, or nerves, and there is abundant evidence from a number of
goats that the cause of bends does not produce such lesions in the
nervous system as are followed by secondary degeneration which can be
revealed by the methods of Marchi or Weigert. The two following
goats may be cited in detail as to this point: in neither was any
degeneration found. Goat XXI (Series II) was used in seven experi-
ments between November 26th and January 18th: it had bends on
December 5th, 11th and 18th (the last being noted as "bad bends"),
and dyspnoea, nearly fatal, on January 2nd: it was killed on January
18th. Goat XV A (Series III) was exposed 27 times between February
2nd and June 10th: it had bends on February 2nd, 20th, 22nd,
March 3rd, 5th, May 13th, loth, 27th and June 6th: it was killed on
June 11th. Thorough examination of the pons, medulla and cord
showed no secondary degeneration in either animal. There are there-
fore reasons for thinking that the cause of bends is peripheral rather
than central. The constant presence of bubbles in the joints has been
already mentioned, and they seem to afford a fairly probable explanation
of most of the cases. Even in those cases in which the muscles are the
seat of pain, it is quite possible that a sensation actually originating in
or around the joints is referred to other parts. The joint pains in man
are often relieved by flexion, and goats evidently try to obtain ease in
the same way (see Plate VI). This fact adds strong confirmation to
the conclusion that the origin of the pain is in or about the joints.
We have already seen that bends, while not the first symptom
to appear as the duration of exposure to high pressure is increased, are
the last symptom to disappear as the decompression is extended, that
bends in short arise in parts of the body which saturate and desaturate
27—2
rather slowly. The synovial fluid satisfies this criterion; on the other
hand the tendons and other dense tissues about the joints are not in
disagreement with it.
Bends occur with a lower degree of supersaturation with air than
any other symptom of compressed-air illness. In goats they are readily
produced after exposures to 30 lbs. or less. The fact that only a
moderate degree of supersaturation is needed to produce them seems
to explain the fact that although they are not the first symptom to
appear as the duration of exposure to very high pressure is increased,
yet a moderate duration of exposure suffices to produce them, in spite
of the fact that they occur in parts of the body with a slow rate of
circulation, as shown by the fact that they are the last symptom to
disappear as the duration of decompression is prolonged.
SUMMARY.
1. The time in which an animal or man exposed to compressed air

becomes saturated with nitrogen varies in different parts of the body
from a few minutes to several hours. The progress of saturation follows
in general the line of a logarithmic curve and is approximately complete
in about five hours in man and in a goat in about three hours.
2. The curve of desaturation after decompression is the same as
that of saturation, provided no bubbles have formed.
3. Those parts of the] body which saturate and desaturate slowly
are of great importance in reference to the production of symptoms after
decompression.
4. No symptoms are produced by rapid decompression from an
excess pressure of 15 pounds, or a little more, to atmospheric pressure,
i.e. from two atmospheres absolute to one. In the same way it is safe
to quickly reduce the absolute pressure to one-half in any part of the
pressure scale up to at least about seven atmospheres: e.g. from six
atmospheres (75 pounds in excess) to three (30 pounds), or from four
atmospheres to two.
5. Decompression is not safe if the pressure of nitrogen inside the
body becomes much more than twice that of the atmospheric nitrogen.
6. In decompressing men or animals from high pressures the first
part should consist in rapidly halving the absolute pressure: subse-
quently the rate of decompression must become slower and slower, so
that the nitrogen pressure in no part of the body ever becomes mor°
than about twice that of the air. A safe rate of decompression can be
calculated with considerable accuracy.
7. Uniform decompression has to be extremely slow to attain the
same results. It fails because it increases the duration of exposure to
high pressure (a great disadvantage in diving work), and makes no use
of the possibility of using a considerable difference in the partial
pressure of nitrogen within and without the body to hasten the desatu-
ration of the tissues. It is needlessly slow at the beginning and usually
dangerously quick near the end.
8. Decompression of men fully saturated at very high pressures
must in any case be of very long duration: and to avoid these long
decompressions the time of exposure to such pressures must be strictly
limited. Tables are given indicating the appropriate mode and dura-
tion of decompression after various periods of exposure at pressures up
to 90 pounds in excess of atmospheric pressure.
9. Numerous experiments on goats and men are detailed in proof
of these principles.
10. The susceptibility of different animals to compressed-air illness
increases in general with their size owing to the corresponding diminu-
tion in their rates of circulation.
11. The average respiratory exchange of goats is about two-thirds
more than that of man; they produce about 0'8 gram, of CO2 per hour
per kilogramme of body weight.
12. The mass of the blood in goats is six and a half or seven and a
half per cent, of the " clean " body weight.
13. The individual variation among goats in their susceptibility
to caisson disease is very large. There is no evidence that this depends
directly on sex, size or blood-volume : there is some evidence that fat-
ness and activity of respiratory exchange are important factors.
14. Death is nearly always due to pulmonary air-embolism, and
paralysis to blockage of vessels in the spinal cord by air. The cause of
" bends " remains undetermined; there are reasons for supposing that
in at least many cases they are due to bubbles in the synovial fluid of
the joints.
15. In our experiments bubbles were found post-mortem most freely
in the blood, fat and synovial fluid; they were not uncommon in the
substance of the spinal cord, but otherwise were very rarely found in the
solid tissues.
APPENDIX I.
Details of the experiments made on Lieutenant Damant and Mr A. Y.
Catto, Gunner, R.N., in the pressure chamber at the Lister Institute.
These experiments were undertaken in July, 1906, as a preliminary
to actual diving experiments in very deep water.
In the first three or four the decompression was controlled from
inside the chamber; in the rest from outside. The subjects remained
closed in the chamber for half an hour after each experiment, the engine
being also kept running so that recompression could be at once begun
if any serious symptom developed. In addition to the actual period of
exposure to each pressure, we have noted the virtual period of exposure
calculated on the assumption that about half the time occupied in com-
pression must be added (see above, p. 362).
In view of the results with goats, the occurrence of decompression
symptoms seemed probable in the more severe experiments. No
symptoms were, however, observed, except considerable itching of the
skin of the fore-arms where it was uncovered. In the compressed air
the well-known alteration in the voice, and corresponding abnormal
sensations about the lips and mouth, were very marked at pressures
exceeding 60 or 70 lbs.
I. July 25th. Actual exposure to 39 lbs. for one hour. Virtual
exposure 69 minutes, decompression in 24 minutes:
Compressed to ... ... ... ... 39 lbs. in 17 minutes.
Waited at 9 ,, for 60
Decompressed to ... 9 ,, in 7
Decompressed to 4 ,, in 1 24.
Decompressed to ... 0 .. in 2
II. July 26th. Actual exposure to 50 lbs., 27 minutes. Virtual
exposure, 39 minutes. Started at 10.37 a.m. Decompression in
34 minutes:
Decompressed to ... 17 „ in 4
Waited at 13 „ for H
Waited at 9 „ for 3
Decompressed to ... 4 >. in 2
Waited at i „ for 8
III. Same day, 3.3 p.m. Exposure to 55 lbs. for 19 minutes.

Virtual 33 minutes. Decompression in 31 minutes:
Compressed to 55 lbs. in 28 minutes.
Decompressed to ... 17 „ in 4 „
13 „ )> 5
9 „ 5
4 „ 10
Decompressed from 4 to 0 .. in 2
The time taken for decompressing from 17 to 13 lbs., &c, was

counted as time at 13 lbs.
IV. July 27th, 10.29 a.m. Exposure to 60 lbs. for 20 minutes.

Virtual exposure 36 minutes. Decompression in 37^ minutes:
Decompressed to ... 22 i. in 5
Waited at 22 » for 5
Waited at 13 „ for Bi
Waited at 9 » for 4
Waited at 4 ,. for 84
Decompressed to ... 0 ,. in 2i
V. Same day, 3.37 p.m. Exposure to 67 lbs. for 18 minutes.

Virtual exposure 36 minutes. Decompression in 36 minutes:
Decompressed to . . 4 „ in 1
Decompressed to ... 0 .. in 3
VI. July 30th, 10.57 a.m. Actual exposure at 74 lbs., 15 minutes.
Virtual exposure 35 minutes. Decompression in 42 minutes:
Compressed to ... ... ... ... 74 lbs. in 39 minutes
Waited at ... ... ... ... 74 ,, for 15
Decompressed to ... ... ... ... 26 „ in 4
Waited at ... ... ... ... 26 „ for 5
Waited at ... ... ... ... 22 ,, for 4
Decompressed to ... ... ... ... 17 „ in
for H
n „
Waited at ... ... ... ... 17 „ „
Decompressed to ... ... ... ... 13 „ in i
Waited at ... ... ... ... 13 „ for 4
Waited at ... ... ... ... 9 „ for 4
Decompressed to ... ... ... ... 4 „ in H ..
Waited at ... ... ... ... 4 „ for 8J „
Decompressed to ... ... ... ... 0 .. in 3
VII. July 31st, 11.0 a.m. Actual exposure to 80 lbs. for 12 minutes.
Virtual exposure, 34 minutes. Decompression in 51 minutes:
Waited at ... ... ... ... 80 ,, for 12
Decompressed to ... ... ... ... 31 ,, in 3
Waited at ... ... ... ... 31 „ for 5
Decompressed to .. ... ... ... 22 „ in 1
Waited at ... ... ... ... 22 „ for 4
Waited at ... ... ... ... 18 „ for 4
Waited at ... ... ... ... 15 „ for 2
Waited at ... ... ... ... 13 „ for 4
Waited at ... ... ... ... 9 „ for 9
Waited at ... ... ... ... 4 „ for 8
Decompressed to ... ... ... ... 0 .. in 3
APPENDIX II».
A DIARY OF THE DEEP DIVING EXPERIMENTS CARRIED OUT OFF
ROTHESAY, ISLE OF BUTE, FROM H.M.S. SPANKER, AUGUST, 1906.
Monday, 20th August.

H.M.S. Spanker arrived at Rothesay about 7 p.m., and was met by
Drs Haldane and Rees and Mr Catto, Gunner, R.N. Arrangements were
made to commence experiments the following day.
Tuesday, 21st August.
All the pumps to be used in the experiments were tested up to a
pressure of 200 feet, and the leakage at this pressure measured. The
pressure gauges, which had been specially graduated for these experi-
ments, were tested and found to give correct readings. The method of
testing employed was to attach the free end of the diving hose to a lead
line, and lower it over the side into the sea to the required depth. The
pumps were then hove round until there was a free supply of air, and
then stopped whilst the reading of the gauge was taken.
The re-compression chamber was tested on the Whitehead torpedo
charging column, and it was found that the pressure could be brought
up to 40 lbs. on the gauge in 3 minutes. There was a leak of 1 lb. per
minute, or, roughly 3 cubic feet. Afterwards Drs Haldane and Rees
were compressed up to about 30 lbs. in order to further test the working
of the chamber.
In the afternoon both divers made a trial dive in 15 fathoms:
Lieutenant Damant Mr Catto
Time of descent 2 minutes 1\ minutes.
,, on bottom 1 hour 1 hour.
„ of ascent 18i minutes 17J minutes
No. 5-minute stops 1 at 30 feet 1 at 20 feet.
„ 10 1 „ 10 „ 1 „ 10 „
Two double pumps were used for each diver in these and the subse-
quent dives. The divers were perfectly comfortable in moving about on
the bottom. It may be mentioned that Lieutenant Damant had not
dived previously beyond about 19 fathoms, and had no experience in
diving except what he had gained in his course of instruction as a
gunnery officer and in experimenting at Portsmouth for the Committee.
Mr Catto had much previous experience in diving work, but had never
dived beyond 23 fathoms.
1
Reprinted from the Report of the Admiralty Committee on Deep Diving, 1907.
Wednesday, 22nd August,
H.M.S. Spanker, off Rothesay.
In the forenoon Mr Catto descended in 23 fathoms, and in the after-
noon Lieutenant Damant did the same :
Mr Catto Lieutenant Damant
Time of descent 2 minutes 2J minutes.
,, on the bottom 20 „ 20 ,,
,, of ascent 35J „ 32J ..
No. 5-minute stops 4 at 50, 40, 30, 20 feet 2 at 50, 40 feet.
„ 10 1 „ 10 feet 2 „ 20, 10 „
Thursday, 23rd August,
H.M.S. Spanker, off Rothesay.
After testing the pumps each diver made a descent to 25 fathoms :
Lieutenant Damant Mr Catto
Time of descent 2 minutes 2 minutes.
„ on the bottom 18| „ 19i „
,, of ascent 37| „ 393 ,,
No. 5-minute stops 3 at 60, 45, 30 feet 3 at 50, 40, 30 feet.
„ 10 2 „ 20, 10 feet 2 „ 20, 10 feet.
Friday, 24th August,
H.M.S. Spanker was taken through the narrows of the Kyles
of Bute and anchored off the entrance of Loch Riddon.
In the morning, after the usual tests had been applied to the pumps,
Mr Catto descended in 27 fathoms, and in the afternoon Lieutenant
Damant went down in a similar depth:
Time of descent 2 minutes ... ... ... 1 minute 20 seconds.
,, on the bottom 16J „ ... ... ... 16| minutes.
,, of ascent 55J ,, ... ... ... 44J ,,
No. 5-minute stops 4 at 60, 50, 40, 30 feet ... ... 4 at 60, 50, 40, 30 feet.
„ 10 ,, „ 1 at 20 feet. (Diver was employed just 2 ,, 20, 10 feet,
under the ship's bottom in examining
a propeller which had been slightly
injured, for 19J minutes before coining
up.)
Saturday, 25th August,
H.M.S. Spanker, off Loch Riddon.
The Spanker shifted her position slightly, and, after the usual
tests of the pumps, both divers descended in 29 fathoms of water :
Time of descent 3 minutes 1J minutes.
„ on the bottom 14J „ 13J „
„ of ascent 46 „ 484 ..
No. 5-minute stops 4 at 70, 50, 40, 30 feet 4 at 66, 54, 40, 30 feet.
No. 10 „ 2 at 20, 10 feet 2 at 20, 10 feet.
Monday 27th August,

Thirty fathoms of water were obtained. Mr Catto was the diver
in the morning. The pumps used were Nos. 3604 and 3593. Six men
were told off for each pump, in reliefs of 5 minutes. Details of the
descent:
Time Remarks
11.22 Glass screwed up. Depth by lead line 304 fathoms.
11.23J Diver under water.
11.234 ,, down 50 feet.
„ 70 „
» HO „
11.24A ,, 150 „
„ ,, 180 ,, ou the bottom. 1 min. 30 sees, in descending. Bevolutions
averaged 32 per min., but fell to 24 for a short time, owing to the great
exertions that were necessary to keep the pumps going at the higher speed.
Diver quite comfortable while moving about on the bottom.
Diver called up.
11.384 „ started up.
„ at 160 feet.
„ „ HO „
11.40J „ >, 120 „
11.41 „ ,, 100 „
11.414 » ,. 70 „ 1st stop. Diver employed in gymnastic exercises. One pump
stopped.
11.464 Diver at 50 feet. 2nd stop.
11.514 „ „ 40 „ 3rd stop.
11.56.J „ „ 30 ,, 4th stop.
12.14 » » 20 „ 5th stop.
12.114 „ „ 10 „ 6th stop. There were no ill-effects. Water jackets gained
20 degrees F.
12.224 Diver called up.
12.23| Glass off.
Afternoon. Lieutenant Damant.

2.144 Screwed up glass.
2.16 „ down 70 feet.
2.164 .. .. 120 »
2.16J „ „ 160 „
2.17 „ ,, 186 ,, on the bottom. 1 minute 20 seconds in descending. Revolu-
tions averaged 30 per minute.
2.29 Diver called up.
2.31 „ at 170 feet.
2.33 „ ,, 120 „ Diver stopped 1J minutes.
2.33^ „ „ 70 ,, 1st stop.
2.384 » » 50 » 2nd "
Time Remarks
2.43J Diver at 40 feet. 3rd stop.
2.484 „ „ 30 ,, 4th „
2.534 „ „ 20 „ 5th „
3.3J .. ,» 10 „ 6th ,,
3.134 „ called up.
3.15J Glass off. There were no ill-effects. Later in the afternoon the pumps were
tested at different temperatures of the water jacket, to see how the leakage
was affected.
Tuesday, 28th August.

In the same locality, Lieutenant Damant made a second descent in
30 fathoms in order to obtain samples of the air in the helmet. The
pumps used were Nos. 3588 and 3592 :
Time Remarks
10.20J ,, on the bottom, 1 minute 40 seconds in going down.
10.34J „ started up.
11.214 Glass off. Whilst on the bottom, diver took two samples whilst at rest. There
was a distinct tide on the bottom, which affected the diver.
Analysis of Samples.
CO2 production in cubic feet
No. of sample CO2 per cent. O2 per cent. per minute
1st •32 20-86 •025
2nd •50 20-43 •041 (? tide)
In the afternoon Mr Catto was in the dress. Pumps Nos. 3588

and 3592 were used:
Time Remarks
2.17 Glass screwed up.
2.18J Diver down 60 feet.
2.19 „ ,, 100 „
2.19| ,, „ 180 ,, on the bottom. The diver took down with him a wire
hawser to shackle on to a sinker.
2.31J Diver called up, but could not come up as he was foul, until—
2.J0J ,, at 140 feet.
2.53 „ 100 „ 1st stop.
2.56 » 80 „ 2nd „
3.1 „ 60 „ 3rd „
3.7 ,, 50 „ 4th „
3.12 „ 40 „ 5th „
3.22 „ 30 „ 6th ,,
3.37 „ 20 „ 7th „
3.52 „ 15 „ 8th „
4.0 „ 10 „ 9th „
4.184 on the surface.
Mr Catto attempted to shackle a hawser on to the sinker. He found

the sinker without the slightest difficulty, and then, having tied his
distance line to it, went back to the hawser. He found this in bights,
and he seems to have got within the coils, and in trying to find the end
of the wire to have fouled his life line. When called up he could not
get away, and it was 20 minutes before he could clear himself. In all
he was down 28f minutes in 30 fathoms of water. The rate of the
pump could not be kept up above 24 revolutions per minute, and the
supply of air was not adequate to his exertions to free himself, so that
he was almost overcome by the excess of CO2. On account of his long
exposure during heavy work, great care was taken in decompressing
him, 1| hours being allowed. There were no ill-effects.
Thursday, 30th August,
Mr Catto made another descent under the same conditions, and
shackled on the hawser to the sinker in 4 minutes after reaching the
bottom. The revolution of the pump averaged 24 to 30 per minute.
The day was very bright, with the sun shining on the water, so that the
diver saw with comparative ease in the water.
In the afternoon Lieutenant Damant, at the same depth, took three
samples of the air in the helmet, and the pumps were tested at 180 feet
pressure. He suffered from no ill-effects:
Time of descent 3 minutes 1 min. 20 sees.
„ on the bottom 12f ,, 13 minutes.
,, of ascent 46£ „ 46J „
No. 5-minute stops 4 at 70, 50, 40, 30 feet 4 at 70, 50, 40, 30 feet.
,,10 ,, ,, 2 at 20, 10 feet 2 at 20, 10 feet.
Analysis of Samples obtained by Lieutenant Damant.

Per cent. First sample Second sample Third sample
C0. 2 ... ... ... -43 -39 -36
O2 :.. ... ... 20-56 20-52 20-47
Deficiency of oxygen ... ... -48 -52 -57
C0. 2 produced in cubic feet per m i n u t e -035 -029 -027
Friday, 31st August,

H.M.S. Spanker moved down to the entrance of Loch Striven;
where 35 fathoms of water could be obtained.
In the morning Lieutenant Damant was the diver. Pumps
Nos. 2593, 3604 and 3592 were tested and used. Six hands were told
off for each pump in reliefs of 5 minutes:
Time Remarks
11.8 Glass screwed up.
11.81 Diver under water.
11.9 „ down 80 feet.
11.91 ., ,, 120 ,,
11.94 » » 150 ,,
11.9} ,, „ 180 „
11.101 >. ,, 200 ,,
11.104 >, >> 216 „ on the bottom. Kevolutions kept at 30 per minute, and
the diver had a good supply of air.
11.134 Diver took samples seated on the shot at the bottom of tlie rope.
11.154 ,, called up.
11.16J „ started up.
11.17 „ at 190 feet.
11.18 „ ,,110 „ Diver stopped to blow off sampling tube.
11.204 >, » 9° ,» 1 s t s t o P-
11.234 „ „ 70 „ 2nd „
11.284 ,, ,, 52 „ 3rd „
11.334 ., 1. 42 „ 4th ,,
11.394 ., „ 32 „ 5th „
11.444 .. •• 22 „ 6th „
11.544 .. .. H ., 7th „
12.44 ,. «alled up.
There was no light on the bottom, which was of soft mud. The
depth by the shot rope was 210 feet. Pressure was 93J'lbs. The
gauge showed a pressure of 216 feet of fresh water with the pumps
stopped, and 220 feet whilst they were heaving. The actual depth, as
carefully measured on the shot rope against the ship's standard measure,
was just over 35 fathoms, 210 feet.
In the afternoon Mr Catto made the same descent, and reached
35 fathoms. He found that the air supply was more than ample. He
walked out to the end of his distance line, and then took a sample of the
air in his helmet:
Time Remarks
2.12 Screwed up glass. Same pumps as last.
2.12| Diver under water.
2.14f „ on the bottom. Revolutions reduced to 24, as the diver found the supply
too much. He proceeded to the end of his distance line before taking his
sample.
2.204 Diver started up.
2.27| „ at 90 feet. 1st stop.
2.30J „ „ 70 „ 2nd „
2.35| „ „ 50 „ 3rd „
2.40| „ „ 40 „ 4th ,,
2.45| „ „ 30 „ 5th ,,
2.50| •„ „ 20 „ 6th „
3.0| „ „ 10 „ 7th „
3.10| „ called up.
Lieut. Damant Mr Catto
CO2 per cent. •14 •53
O2 „ „ 20-89 2034
Deficiency of 0.2 per cent •15 •70
Monday, 3rd September.

Experiments on rest and measured work were carried out, by means
of an arrangement of rope and pulleys by which the diver on the bottom
raised and lowered a 56 lb. weight suspended in view of those on deck.
The heavy rope and blocks used caused great friction and resistance.
Time Remarks
2.26 Diver, Mr Catto, descended.
2.27J ,, on bottom, 142 feet.
2.31 ,, took sample sitting on the shot. (No. 1.)
Two pumps at 30 revolutions per minute.
Raised the weight 4 times 5 feet, at the rate of one lift per minute.
2.36 „ took sample. (No. 2.)
Raised weight 7 times 5 feet in 5^ minutes.
2.42 „ took sample. (No. 3.)
2.45 ,, started up.
3.23 ,, on surface, no ill-effects.
3.3J ,, Lieutenant Damant, started down.

3.4J ,, down 100 feet.
3.5 „ on bottom, 139 feet.
4.0 „ took sample sitting on the shot. . (No. 4.)
Two pumps at 26 revolutions.
Raised weight 5 times in If minutes.
4.3 ,, took sample. (No. 5.)
Raised weight 3 feet 18 times in 6J minutes.
4.10 ,, took sample. Pump 24 revolutions. (No. 6.)
4.13J ,, started up.
4.52J ,, at the surface. No ill-effects.
CO2 •30Percent.|MrCatto S a m p l e No.
Oa 20
CO 2 ... ... -70 No. 2.
O, ... ... 20-29
CO3 -71 „ ) N
20-23
- f " "
CO2 ... ... -18 Lieutenant Damant. Sample No. 4.
O, ... ... 20-73
73 I „ _
20-12 | No. 5.
No 6
20-36 " ^ " " - -
Tuesday, 4th September.
The Spanker was anchored in six fathoms of water, and experiments
were made on the bottom by Dr Haldane, Lieutenant Damant, and
Mr Catto on the risks of blowing up. After being compressed in the air
chamber to teach them to open their Eustachian tubes, Lieutenant
and Commander E. V. F. R. Dugmore, Lieutenant G. N. Henson,
Jack Haldane (age 13) all made descents in six fathoms of water. This
was the first time that these had ever dived in a diving dress, which
illustrates the usefulness of the re-compression chamber in the practical
teaching of divers.
Wednesday, 5th September.

Exhaustive tests were made as to the leakage of the pumps and
composition of the air, with the water jackets at various temperatures.
The results are embodied in the Report. These experiments concluded
the work undertaken for the Committee.
APPENDIX III.
We give here some illustrative protocols of certain important animal

experiments.
1. Comparison of stage (93 minutes) and uniform (100 minutes)
decompression after 2 hours exposure at 1b lbs.
(a) 15.3.07. Goats 3, 4, X A, XI A, XVI A, XIX A, XXV A.
Started up 10.23
Beached 75 lbs. 11.3
Started from 75 lbs. 12.43
31£ 12.49
27 12.59
22 1.14
18 1.29 Stage decompression.
13J 1.44 Total = 93 mins.
1.59
2.14
Reached 0 lbs. 2.16
XIX A had bends left hind-leg at 2.20. X A passed urine not frothy
at 2.27. Rest nil.
A E. BOYCOTT, G. C. (J. DAMANT AND J. S. HALDANB 437
(6) 18.3.07. Goats XII A, XIII A, XV A, XVIII A, XX A,

XXI A, XXIII A.
Started up 10.53
Reached 75 lbs. 11.35
31J .. 1.20
27 „ 1.30
22 „ 1.45
18 „ 2.0 Stage decompression.
134 ,, 2.15 Total = 93 mins.
!) •• 2.30
2.45
Reached 0 lbs. 2.47 J
XX A urine no froth at 2.54. XIIIA at 2.55 seemed uneasy in hind-

legs and lay down but nothing definite; at 3.0 right hind-leg slight
limp and foot-drop ; foot-drop very marked at 3.10 and could hardly
walk ; alright at 4.0. Rest nil.
(c) 14. 3. 07. Goats XXIV A, XXVI A, XXVII A, XXVIII A,
XXIX A.
Started up 10.35
Reached ! 75 lbs. ... 11.14
s> 314 ,. 1.1
27 „ 1.11
224 „ 1.26
,, 18~ „ 1.41 - Stage decompression.
>) 134 ,, 1.56 Total = 93 mins.
,, 9 „ 2.11
j t 44 „ 2.26
Reached . Olbs. 2.28^
XXIV A showed bends left fore-leg during decompression from 4J-

to 0; XXVIIA passed urine no froth at 2.32; XXVIII A bends left
hind-leg at 2.40. Rest nil.
(d) 12.3.07. Goats XXIV A, XXVI A, XXVII A, XXVIII A,
XXIX A.
Started up 12.53
Reached 75 lbs. 1.1
Started from 75 lbs. 3.0 .
Reached 60 lbs. ... 3.20
45 „ 3.40 (_ Uniform decompression.
30 „ 4.0 I Total = 100 mins.
15 „ 4.20
0 „ 4.40 •
XXIX A bleated at about 1 lb., gnawed side, lay down; both fore-legs
completely paralysed and hind-legs unsteady; kept head bent round on
left side; bleated continually; no dyspnoea till 4.55, when it was moderate;
seemed like to die. At 5.10 no dyspnoea, stopped bleating, could just
stand. Could walk at 5.35. XXVIIA passed very frothy urine at 4.41;
both hind-legs bad bends at 4.50, right fore-leg at 5.10. XXIV A left
hind-leg partial foot-drop and bends at 4.50; left fore-leg bends at 5.0.
XXVIIIA bad bends right fore-leg, won't stand up; at 5.0 could not
stand, had constant nystagmus, bleated; at 5.10 left hind-leg bad bends,
nystagmus stopped, no bleating ; walked very badly at 5.30. XXVIA
no symptoms.
(e) 19.3. 07. Goats 3, 4, X A, XI A, XVI A, XIX A, XXV A.
Started up ... ... ... 10.30
Beached 75 lbs. ... ... ... 11.14
Started from 75 lbs. ... ... 12.52
Beached 60 lbs. ... ... ... 1.12
,, 45 „ ... ... ... 1.32 Uniform decompression.
„ 30 „ ... ... ... 1.52 j Total = 100 mins.
„ 15 „ ... ... ... 2.12 j
0 „ ... ... ... 2.32'
XVIA came out with bad bends left hind and right fore-legs; could
hardly walk and kept head twisted round to left; much better at 2.50.
XIX A urine at 2.34 full of froth ; bends right fore-leg. X A began
bleating at 2.38 but showed nothing till 2.44 when he had complete
foot-drop right fore-leg and bends left hind-leg; at 2.50 right fore-leg
paralysed, could not stand up, left fore-leg also weak; urine at 2.50 a
little froth. XXV A cried out a bit, belly very tight, refuses to move,
evidently far from well: died between 8 and 8.30 a.m. next day: a good
many bubbles in right heart. Rest nil.
(/) 20.3.07. Goats XII A, XIII A, XV A, XVIII A, XX A,
XXI A, XXIII A.
Started up ... ... ... 11.5
Beached 75 lbs. ... ... ... 11.47
Started from 75 lbs. ... ... 1.26 \
Beached 60 lbs. ... ... ... 1.46 |
,, 45 „ ... ... ... 2.6 I Uniform decompression.
30 „ ... ... ... 2.26 [ Total = 100 mins.
15 „ ... ... ... 2.46
0 ,. ... ... ... 3.6 /
A goat unknown aborted two foetuses 2 in. long; they were quite
warm when found, so probably during decompression. XXIIIA very
A. E. BOYCOTT, G-. C. C. DAMANT AND J. S. HALDANB 439
frothy urine at 3.8. XIIIA dyspnoea, both hind-legs dragging; at 3.25

lying down, moaning bleat, tongue and lips getting cyanosed, dyspnoea
not violent. Made sure it was going to die, but at 3.55 it got up and
showed only bends right fore-leg and weakness in both hind-legs. At
5.0 seemed all right. Rest nil.
(2) The effects of a sudden drop of 51 lbs. in different parts of the

scale of absolute pressure.
(a) 26.3.07. Goats XXIVA, XXVIA, XXVIIA, XXVIIIA,
XXIX A.
Started up .. ... ... 10.0
Reached 75 lbs. ... ... ... 10.46
Left 75 lbs. ... ... ... 1.23
Reached 24 lbs. ... ... ... 1.24J
Left 24 lbs. ... ... ... 2.25
„ 14 ,, ... ... ... 2.55
„ 8 „ ... ... ... 3.18J
Reached 0 lbs. ... ... ... 3.21J Total = 118J mins.
No symptom's during decompression. XXVIII A passed frothy urine

at 3.23; at 3.31 had bad bends, evidently very uneasy generally; better
at 4.0. XXIX A urine no froth at 3.31. Rest nil. (The immediate object
of the experiment having been attained, an unwise quickening of the
end of decompression gave XXVIIIA bad bends.)
(b) 23 . 5.07. Goats 7, 9, XXX A, XXXII A, XXVII.
Start up ... ... 9.55
Reach 75 lbs. ... ... ... 10.35
Left 75 lbs. ... ... ... 1.15
Reach 24 lbs. ... ... ... 1.15.40"
Left 24 lbs. ... ... ... 2.15
„ 14 „ .- 2.45
„ 8 ., - 8.6
,, 4 ,, ... ... ... 3.25 Total = 131 mins.
No symptoms during decompression. 7 limped right hind-leg

on coming out; urine 3.30 no froth. 9 right hind-leg bends at 3.35.
XXXIIA urine 3.35 no froth. Rest nil.
(c) 27.3.07. Goats XXIV A, XXVI A, XXVII A, XXVIII A,
XXIX A.
Started up ... ... ... ... 11.0
Reached 51 lbs. ... ... ... ... 11.30
Left 51 lbs. ... ... ... ... 2.16
Reached 0 lbs. ... ... ... ... 2.20J
XXVIIIA very unsteady on hind-legs at 2.25, passed urine fall of
froth; legs gave way; at 2.30 lying down grunting, constant nystagmus,
mucous membranes not pale ; at 2.38 respiratory movements almost
stopped; died 2.44 p.m. Ordinary moderate bubbling. XXVIIA
bends left fore-leg at 2.27, bad; a little left next day. XXIV A left
hind-leg bends at 2.29; had pretty marked dyspnoea at 2.40. XXIX A
urine 2.40 no froth, seemed uneasy, kept lying down but could make
out nothing definite. XXVIA no symptoms.
(d) 24.5 .07. Goats 7, 9, XXVII A, XXX A, XXXII A, XXVII.

Started up ... ... ... ... 9.51
Beached 51 lbs. ... ... ... ... 10.15
Started down ... ... ... ... 1.3
Reached Olbs. ... ... ... ... 1.7
XXX A, urine 1.10 much froth, no symptoms. XXVII A, bad bends

left fore-leg, jumpy hind-legs. 7, bends right fore-leg. 9, bends
right fore-leg; slight dyspnoea, bends bad, both hind-legs wobbly;
dyspnoea gone by 1.40 and legs alright. XXXII A bleating, won't
stand up, dyspnoea; died 1.34 p.m. Bad general bubbling. XXVII
bends right fore-leg.
(e) 5 . 6 . 07. Goats XII A, XVI A, XXIII A.

Started up ... ... ... ... 9.52
Reached 45 lbs. ... ... ... ... 10.H
Left 45 lbs. ... ... ... ... 12.3
Reached Olbs. ... ... ... ... 12.3.33"
„ -61bs. ... ... ... ... 12.10
XVIA uneasy at —5 lbs., paraplegic at 12.10, struggling and bleating,

dyspnoea. XII A bends right fore-leg 12.19, bleating at 12.28.
XXIII A tried to get up at 12.25 but failed once; then got up, right
hind-leg paralysed; both gone just afterwards, could just crawl across
tank; dyspnoea at 12.28. Raised pressure to atmospheric and opened
tank at 1.10. X I I A got up and seemed alright. XXIIIA and XVIA
lay log-like, conscious, breathing slightly and slowly. At 1.40 XXIIIA
could rest on fore-legs, hind-legs completely paralysed, ate hay; seemed
pretty well except for paraplegia at 4.0. (Condition did not improve
and it was killed six days later.) XVI A died at 3.20 p.m. A few
small bubbles in right auricle and right femoral vein.
(b) 1 2 . 6 . 07. Goats 7, XXIV A, XXIX A, XXX A.

Started up ... ... ... ... 9.55
Beached 39 lbs. ... ... ... ... 10.13
Left 39 lbs. ... ... ... ... 12.4.20"
Reached Olbs. ... ... ... ... 12.4.50"
LeftOlbs. ... ... ... ... 12.6.10"
Beached - O l b s . ... ... ... ... 12.11.5"
XXX A bends right fore-leg at 12.13, dyspnoea at 12.30. XXIX A

bends right fore-leg at 12.14, lay down, dyspnoea at 12.24. 7 lay
down, dyspnoea at 12.14. XXIV A no symptoms. Raised pressure to
normal and opened up at 1.15. 7 showed bends left fore-leg and
had slight dyspnoea. XXX A seemed alright. XXIV A and XXIX A
were very quiet but no definite symptoms. All alright at 3.30.
APPENDIX IV.
TABLE I.
Stoppages during the ascent of a diver after ordinary limits of time from surface.
Pressure
Depth Time from surface Approximate Stoppages in minutes at different depths* Total time
Pounds per to beginning time to ^ * N for ascent
Feet Fathoms square inch of ascent first stop 60 ft. 50 ft. 40 ft. 30 ft. 20 ft. 10 ft. in mins.
0-36 0-6 0-16 No limit 0-1
36-42 6-7 16-18J Over 3 hours 1 — _ _ _ _ 5 6
Up to 1 hour
42-48 7-8 18J-21 1-3 hours
Over 3 hours
n —
—
—
—
—
—
—
—
—
—
5
10
4
Up to 4 hour Hi
4-14 hours ... 2 — — — — — 5 2
48-54 8-9 21-24 14-3 hours ... 2 — — — — _ 10 7
Over 3 hours 2 12
Up to 20 mins. _ _ _ _ _ _ 20
22
20-45 mins. ... 2 — _ _ _ _ 5 2
54-60 9-10 24-26J |-14 hours ... 2 — — — — — 10 7
14-3 hours ... 2 — — — — 5 15 12
Over 3 hours 2 — — — — 10 20 22
Up to I hour 2 32
\-\ hour 2 5 2
60-66 10-11 26i-29J J-l hour 2 10 7
1-2 hours 2 15 15
2-3 hours ... 2 — 10 20 22
Up to £ hour 32
I
2 2
J-4hour 2 — 3 5 4
66-72 11-12 291-32
4-1 hour 5 12 10
2
1-2 hours ... 2 — 10 20 19
(Up to 20 mins. 2 32
72-78 12-13 32-341 j20-45 mins. ... 2 10 7
(f-li hours ... 2 — — 10 20 17
(Up to 20 mins. 2 5 32
7
78-84 13-14 341-37 \ 20-45 mins. ... 2 — — 5 15 22
(f-l| hours ... 2 — — 10 20
I Up to 10 mins. 2 32
3 5
J 10-20 mins. ... 2 3 5
84-90 14-15 37-40 120-40 mins. ... 2 10
5 15 22
(40-60 mins. ... 2 — — — 3 10 15
3 30
Up to 10 mins.
I
3 6
10-20 mins. ... 2 3 5
90-96 15-16 40-424 2 10
20-35 mins. ... 5 15 22
35-55 mins. ... 2 10 15
3 30
(Up to 15 mins. 3 5 11
96-108 16-18 42*-48 < 15-30 mins. ... 3 — — — 3 7 10
3 23
130-40 mins. ... — — — 5 10 15 33
(Upp to 15 mins. 3 — — — 2 3 7
3 15
108-120 18-20 48-531 15-25
ire mins. ...
3
— — — 5 5 10 23
25-35 mins. ... 10 15 33
Dp to 15 mins. 3 5 7
3 17
120-132 20-22 534-59 15-30 mins. ... 10 15 33
Up to 12 mins. 3 5 5
59-64J 112-25 mins. ... 3 16
132-144 22-24 12
3 — 2 10 32
jUp to 10 mins. 5 5 16
144-156 24-26 64J-70 110-20 mins. ... 3
3 10 12 32
J Up to 10 mins. 5 5 18
156-168 26-28 70-75 (10-16 mins. ... 3
3 — 2 7 10 30
j Up to 9 mins. 5 5 18
168-180 28-30 75-80J (9—14 mins. ... 3
3 — 2 7 10 30
180-192 30-32 80J-86
Up to 13 mins. — 2 7 10 30
Up to 12 mins. 3
192-204 32-34 86-914 2 2 7 10 32
During each stoppage the diver should continue to move his arms and legs.
TABLE II.
Stoppages during the ascent of a diver after delay beyond the ordinary
limits of lime from surface.
Pressure Time from Total
Depth surface to Approximate Stoppages in minutes at different depth; time for
\\atrtTI riinfr
Feet Fathoms square inch UGgilllllllg Lirn6 to / in mins.

of ascent first stop 80 ft. 70 ft. 60 ft. 50 ft. 40 ft. 30 ft. 20 ft. lOftT
60-66 10-11 261-294 Over 3 hours 2 _ _ _ 10 30 42
66-72 11-12 29J-32 -(2-3 hours ... 2 — — — — 10 30 42
(Over 3 hours 2 — 20 30 52
72-78 12-13 32-344 -114-21
! hours 2 — — — — — — 20 25 47
1 Over 24 hours 2 — 30 30 62
IflJ-2 hours ... 2 — 15 30 47
78-84 13-14 34J-37 - 2-3 hours ... 2 — 5 30 30 67
\ Over 3 hours 2 — 10 30 35 77
fl-14 hours ... 2 — 5 15 25 47
84-90 14-15 37-40 < 14-24 hours... 2 — 5 30 35 72
1[Over"24 hours 2 — 20 35 35 92
1f l - H hours ... 2 5 15 30 52
90-96 15-16 40-424 H14-24 hours... 2 — „
10 30 35 77
1[ Over 2i hours 2 — 30 35 35 102
40-60 minutes 2 — _ 10 15 20 47
96-108 16-18 421-48 - 1-2 hours ... 2 5 15 25 35 82
{Over 2 hours 2 — 15 30 35 40 122
35-60 minutes 2 — 5 10 15 25 57
108-120 18-20 48-53J " 1-2 hours ... 2 — 10 20 30 35 97
1[ Over 2 hours 2 — 30 35 35 40 142
| 4~i hours ... 3 — 5 10 15 20 53
120-132 20-22 534-59 i f-14 hours ... 3 — 5 10 20 30 30 98
1I Over l i hours 3 — 15 30 35 40 40 163
|'25-45 minutes 3 — 3 5 10 15 25 61
132-144 22-24 59-641 J-14 hours ... 3 — 10 10 20 30 35 108
, Over H hours 3 — — 30 30 35 40 40 178
1
f 20-35 minutes 3 — 3 5 10 15 20 56
144-156 24-26 64A-70 •! 35-60 minutes 3 — 7 10 15 30 30 95
I,Over 1 hour
'16-30 minutes
3
3
—
—
20 25 30
3 5
35
10
40
15 20
40 193
56
156-168 26-28 70-75 | 4-1 hour 3 — 3 10 10 15 30 30 101
, Over 1 hour g 5 25 25 30 35 40 40 203
r 1
14-20minutes 3 — 3 3 7 10 15 41
168-182 28-30
|
75-804 j 20-30 minutes 3 — 2 2 3 10 15 25 60
4-1 hour ... 3 — 3 3 7 10 20 30 35 111
iOver 1 hour 3 — 15 25 30 30 35 40 40 218
I|
[13-20 minutes 3 — 3 3 7 15 15 46
182-194 30-32 804-86 j 20-30 minutes 3 — 3 3 5 10 15 25 64
4-1 hour ... 3 — 3 5 10 12 20 30 35 118
|, Over 1 hour 3 5 20 25 30 30 35 40 40 228
jfl2-20minutes 3 — 3 3 5 7 10 20 51
194-206 32-34 86-914 j 20-30 minutes 3 — 3 3 3 5 10 20 20 67
4-1 hour 3 3 3 5 10 15 20 30 35 124
3 15
I, Over 1 hour 20 25 30 30 35 40 40 238

1908 J Hygiene - The Prevention of Compressed-Air Illness (Boycott)

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1908 J Hygiene - The Prevention of Compressed-Air Illness (Boycott)

Uploaded by

Copyright:

Available Formats

342

THE PEEVENTION OF COMPRESSED-AIR ILLNESS.

AND J. S. HALDANE, M.D., F.R.S.

{From the Lister Institute of Preventive Medicine.}

[With 7 Figures and 3 Plates.]

MEN who have been working in compressed air, as in diving, pre-

A. The rate of saturation of the body with nitrogen during

after 23 rounds of the circulation, three-fourths after 46 rounds, seven-

Experiments on animals have shown that the venous blood entering

These results seem to show conclusively that the kidney substance

B. The rate of desaturation of the body with nitrogen during

will further reduce it by 28% of 4-5-3, i.e. by 042 atmosphere.

Inspection of Fig. 2 shows that with uniform decompression the

during decompression. The rate of desaturation will thus be increased

C. The limits of safety in decompression.

D. Practical measures for avoiding Compressed-air Illness.

From the foregoing discussion the general nature of the measures

(1) Diving work.

In descending or ascending a diver usually makes use of a rope

calculation. The possible occurrence of slight symptoms after surface

a moderate degree the length of uniform decompression. It must be

Very prolonged uniform decompressions are extremely tedious, and it

divers after exposure for varying periods of time at different depths.

at 25 fathoms. The diver was, however, a man of heavy build with

(2) Work in caissons, tunnels, and diving bells.

PART II. EXPERIMENTAL.

2. Choice of earperimental animals.

invariably fatal to goats. In the same way dogs, with a respiratory

3. Respiratory exchange of goats.

The observations were made by using the pressure-box as a respira-

4. Method of conducting the experiments.

' Bends " of fore-leg in a goat.

some pains to secure that the deaths should be as few as possible so

5. The symptoms observed in goats.

"Bends" maybe seen immediately at, or indeed (but very rarely)

the general experience of caisson works as well as our own experiments

n minutes.. 2 2 30 30 11 10 uni- 30 52 30 l 10 30 10 uni- 30 10 uni-

Exposure minutes 3 10 30 60 120 180 180 120 120 240 240

6. Results of goat experiments.

TABLE IX. Showing the decompressions of goats from, 75 /6s. and

2 r 3£ .rfko « to © *>• o c ^ os to ui © oc iSo. of goats P

These experiments show that goats have taken up enough air in

exposures and rapid decompressions such symptoms as appear are more

These figures show that with a duration of exposure up to about

A sudden drop of about 50 lbs. from 75 lbs. positive to 27 or 24 lbs.

(V) Experiments showing the importance of the mode and spacing

TABLE XVI. Showing the comparison between the results of stage

B. Experiments on identical animals.

(VI) Experiments illustrating the difference between different

bubbled; it is interesting to note that there were no bubbles in the

7. Individual variation among the experimental animals in their

In all, therefore, goats X and XIII showed mild symptoms three

Series I, of 5 deaths, 3 were in females, a distribution of fatalities

Series HI and IV:

Influence of size. In the same way the influence of size on sus-

8. The pathology of caisson disease in goats.

These goats were killed before the expiration of the appropriate

2nd cervical. 3rd dorsal.

Fig. 7 shows the distribution of bubbles in another case: note the

an extent that their removal becomes impossible. It is also clear that

These observations seem to be fairly concordant. In 15 minutes the

A histological point of some practical importance arises in connection

1. The time in which an animal or man exposed to compressed air